AI Data Infrastructure for the Global South with Kate Kallott

Hi, I'm Scott Hansselman. This is Hi, I'm Scott Hansselman. This is another episode of Hansel Minutes in another episode of Hansel Minutes in another episode of Hansel Minutes in association with the ACM Bitecast. Today association with the ACM Bitecast. Today association with the ACM Bitecast. Today I'm chatting with founder and CEO of I'm chatting with founder and CEO of I'm chatting with founder and CEO of Amini, Kate Ko. How are you? I'm good. Amini, Kate Ko. How are you? I'm good. Amini, Kate Ko. How are you? I'm good. How are you? I'm very well. Thank you How are you? I'm very well. Thank you How are you? I'm very well. Thank you very much for spending time with us very much for spending time with us very much for spending time with us today. So you're joining us from your today. So you're joining us from your today. So you're joining us from your offices in Nairobi. Is that right? offices in Nairobi. Is that right? offices in Nairobi. Is that right? Indeed. Indeed. Here in Nairobi, Kenya. Indeed. Indeed. Here in Nairobi, Kenya. Indeed. Indeed. Here in Nairobi, Kenya. Yeah. And Amini.ai is the website. And Yeah. And Amini.ai is the website. And Yeah. And Amini.ai is the website. And one of the things that you come right one of the things that you come right one of the things that you come right out and say on the homepage uh you say out and say on the homepage uh you say out and say on the homepage uh you say data is the biggest barrier to AI data is the biggest barrier to AI data is the biggest barrier to AI adoption in the global south. That is adoption in the global south. That is adoption in the global south. That is the the uh the thesis here that we're the the uh the thesis here that we're the the uh the thesis here that we're working on. Why why are we missing data? working on. Why why are we missing data? working on. Why why are we missing data? Why is the global south data scarce? Why is the global south data scarce? Why is the global south data scarce? One of the things you have to remember One of the things you have to remember One of the things you have to remember is that when you look at um the digital is that when you look at um the digital is that when you look at um the digital divide, you still have 2.6 billion divide, you still have 2.6 billion divide, you still have 2.6 billion people that are unconnected in the people that are unconnected in the people that are unconnected in the world. And basically unconnected means world. And basically unconnected means world. And basically unconnected means invisible. And most of those people are invisible. And most of those people are invisible. And most of those people are based in the global south, which means based in the global south, which means based in the global south, which means they don't have internet connectivity.

they don't have internet connectivity. they don't have internet connectivity. They're not online. So they're not They're not online. So they're not They're not online. So they're not represented by the data sets that are represented by the data sets that are represented by the data sets that are being generated today um to basically being generated today um to basically being generated today um to basically train AI models. So when I moved from train AI models. So when I moved from train AI models. So when I moved from New York City to to Kenya, I just had an New York City to to Kenya, I just had an New York City to to Kenya, I just had an idea with my team which was to build an idea with my team which was to build an idea with my team which was to build an AI company for Africa. But then as data AI company for Africa. But then as data AI company for Africa. But then as data scientists, we start building and you scientists, we start building and you scientists, we start building and you realize well there is no data. So what realize well there is no data. So what realize well there is no data. So what do we do now? Um so we we thought that do we do now? Um so we we thought that do we do now? Um so we we thought that if we were running into that issue, we if we were running into that issue, we if we were running into that issue, we were actually the rest of the ecosystem were actually the rest of the ecosystem were actually the rest of the ecosystem as well. So we decided to pivot towards as well. So we decided to pivot towards as well. So we decided to pivot towards building the data infrastructure for building the data infrastructure for building the data infrastructure for Africa and the global south. Africa and the global south. Africa and the global south. So we know that the corpus that AIs are So we know that the corpus that AIs are So we know that the corpus that AIs are trained on is wide and vast, but it's trained on is wide and vast, but it's trained on is wide and vast, but it's also bi biased towards English. It's also bi biased towards English. It's also bi biased towards English. It's biased towards uh the northern biased towards uh the northern biased towards uh the northern hemisphere. Uh what kind of data are we hemisphere. Uh what kind of data are we hemisphere. Uh what kind of data are we missing? Cuz I can think of what I think missing? Cuz I can think of what I think missing? Cuz I can think of what I think the I can guess what I think some of the the I can guess what I think some of the the I can guess what I think some of the data is depending on whether you're data is depending on whether you're data is depending on whether you're talking about data for a GPT, data for a talking about data for a GPT, data for a talking about data for a GPT, data for a chatbot, or whether it be uh more chatbot, or whether it be uh more chatbot, or whether it be uh more specific data. It's actually affecting specific data. It's actually affecting specific data. It's actually affecting many different types of data. So you many different types of data. So you many different types of data. So you would think about data in terms of would think about data in terms of would think about data in terms of languages. For example, you have more languages. For example, you have more languages. For example, you have more than 2,500 languages in Africa. A lot of than 2,500 languages in Africa. A lot of than 2,500 languages in Africa. A lot of those are only spoken, they're not even those are only spoken, they're not even those are only spoken, they're not even written, right? So how are you going to written, right? So how are you going to written, right? So how are you going to go and find the data sets to be able to go and find the data sets to be able to go and find the data sets to be able to capture those languages, those dialects capture those languages, those dialects capture those languages, those dialects that are only spoken by very very few that are only spoken by very very few that are only spoken by very very few communities across the continent. The communities across the continent. The communities across the continent. The second one would be data in terms of um second one would be data in terms of um second one would be data in terms of um environmental data, which is the first environmental data, which is the first environmental data, which is the first one, the first big big big challenge one, the first big big big challenge one, the first big big big challenge that we decided to tackle. Um when you

that we decided to tackle. Um when you that we decided to tackle. Um when you think about for for example satellite think about for for example satellite think about for for example satellite constellation geospatial data we have constellation geospatial data we have constellation geospatial data we have abundance of data in the global north abundance of data in the global north abundance of data in the global north you know in the US in Europe you're able you know in the US in Europe you're able you know in the US in Europe you're able to access weather forecast very easily to access weather forecast very easily to access weather forecast very easily on your phone um but when it comes to on your phone um but when it comes to on your phone um but when it comes to Africa actually the continent has only Africa actually the continent has only Africa actually the continent has only 1/8 of the minimum um density required 1/8 of the minimum um density required 1/8 of the minimum um density required of meteorological station so how are you of meteorological station so how are you of meteorological station so how are you even able to predict weather and even able to predict weather and even able to predict weather and understand when there will be an extreme understand when there will be an extreme understand when there will be an extreme rainfall when there will be floods when rainfall when there will be floods when rainfall when there will be floods when there will be drought it makes it very there will be drought it makes it very there will be drought it makes it very very difficult to access high quality um very difficult to access high quality um very difficult to access high quality um weather data on the continent. So when weather data on the continent. So when weather data on the continent. So when you think about that data scarcity is you think about that data scarcity is you think about that data scarcity is not just one modality is actually not just one modality is actually not just one modality is actually multimodality and multimodality and multimodality and multi-industries. Think even about multi-industries. Think even about multi-industries. Think even about culture. A lot of our culture is not culture. A lot of our culture is not culture. A lot of our culture is not online neither. It's very very difficult online neither. It's very very difficult online neither. It's very very difficult to um digitize that culture and to um digitize that culture and to um digitize that culture and understand um how the continent has understand um how the continent has understand um how the continent has evolved over the past couple of decades. evolved over the past couple of decades. evolved over the past couple of decades. So we've decided to focus on So we've decided to focus on So we've decided to focus on environmental but it is it is a environmental but it is it is a environmental but it is it is a challenge that's affecting many challenge that's affecting many challenge that's affecting many different types of data and there is a different types of data and there is a different types of data and there is a big gap to close here. So when you say big gap to close here. So when you say big gap to close here. So when you say focus on environmental are you is the focus on environmental are you is the focus on environmental are you is the issue uh satellite coverage is it is it issue uh satellite coverage is it is it issue uh satellite coverage is it is it do we solve that with drones flying over do we solve that with drones flying over do we solve that with drones flying over hectares of land like how do you get hectares of land like how do you get hectares of land like how do you get this data because you're you're coming this data because you're you're coming this data because you're you're coming from behind in this race. Yeah, you from behind in this race. Yeah, you from behind in this race. Yeah, you know, it would be very expensive to fly know, it would be very expensive to fly know, it would be very expensive to fly drones all over Africa. So that would be drones all over Africa. So that would be drones all over Africa. So that would be uh quite impossible. But um one of the uh quite impossible. But um one of the uh quite impossible. But um one of the things to to mention is that we're using

things to to mention is that we're using things to to mention is that we're using satellite imagery. So that's the first satellite imagery. So that's the first satellite imagery. So that's the first data modality that we started to work data modality that we started to work data modality that we started to work with. Um and what we realized is that a with. Um and what we realized is that a with. Um and what we realized is that a lot of the constellations, the satellite lot of the constellations, the satellite lot of the constellations, the satellite constellations that have been deployed constellations that have been deployed constellations that have been deployed um um in the past are actually when you um um in the past are actually when you um um in the past are actually when you start looking down to Africa and you start looking down to Africa and you start looking down to Africa and you start going into areas that are less start going into areas that are less start going into areas that are less densely populated, Africa becomes dark. densely populated, Africa becomes dark. densely populated, Africa becomes dark. So there is a lot of work that needs to So there is a lot of work that needs to So there is a lot of work that needs to happen where you couple geospatial happen where you couple geospatial happen where you couple geospatial imagery with machine learning and AI to imagery with machine learning and AI to imagery with machine learning and AI to be able to really truly understand and be able to really truly understand and be able to really truly understand and extract insights out of this imagery. So extract insights out of this imagery. So extract insights out of this imagery. So we're using satellite as a satellites or we're using satellite as a satellites or we're using satellite as a satellites or geospatial data as the first modality geospatial data as the first modality geospatial data as the first modality but then we also working with companies but then we also working with companies but then we also working with companies startups private sector companies who startups private sector companies who startups private sector companies who have a lot of data you know because have a lot of data you know because have a lot of data you know because we're talking about data scarcity but it we're talking about data scarcity but it we're talking about data scarcity but it is also information in accessibility. is also information in accessibility. is also information in accessibility. Sometimes the data is there but people Sometimes the data is there but people Sometimes the data is there but people don't know what to do with it. So we don't know what to do with it. So we don't know what to do with it. So we realized that we needed to put a lot of realized that we needed to put a lot of realized that we needed to put a lot of effort um and made a conscious choice in effort um and made a conscious choice in effort um and made a conscious choice in developing our own ecosystem of data developing our own ecosystem of data developing our own ecosystem of data partners who are feeding us data. the partners who are feeding us data. the partners who are feeding us data. the data alone wouldn't mean anything but data alone wouldn't mean anything but data alone wouldn't mean anything but the data coupled with other types of the data coupled with other types of the data coupled with other types of data geospatial data that's when it data geospatial data that's when it data geospatial data that's when it becomes very very interesting to fusion becomes very very interesting to fusion becomes very very interesting to fusion all these different data types and see all these different data types and see all these different data types and see what are the source of insights that we what are the source of insights that we what are the source of insights that we can extract and you call out uh what you can extract and you call out uh what you can extract and you call out uh what you call ground truth sources and can you call ground truth sources and can you call ground truth sources and can you explain that that you get a bunch of explain that that you get a bunch of explain that that you get a bunch of data but you don't know if it's true or data but you don't know if it's true or data but you don't know if it's true or not what is a ground truth source and not what is a ground truth source and not what is a ground truth source and why is that alignment so important so why is that alignment so important so why is that alignment so important so when you look at satellite imagery it's

when you look at satellite imagery it's when you look at satellite imagery it's extremely biased, right? Is the reality extremely biased, right? Is the reality extremely biased, right? Is the reality of the sky. If you look down um and you of the sky. If you look down um and you of the sky. If you look down um and you look at the ground, usually there's some look at the ground, usually there's some look at the ground, usually there's some sort of calibration and validation that sort of calibration and validation that sort of calibration and validation that needs to happen between what you can see needs to happen between what you can see needs to happen between what you can see from the sky and what you can actually from the sky and what you can actually from the sky and what you can actually measure from the ground. So you actually measure from the ground. So you actually measure from the ground. So you actually need to have just like in computer need to have just like in computer need to have just like in computer vision, you need to have ground truth vision, you need to have ground truth vision, you need to have ground truth when you work with satellite imagery. So when you work with satellite imagery. So when you work with satellite imagery. So what we do is that a lot of our data what we do is that a lot of our data what we do is that a lot of our data partners would go and for example partners would go and for example partners would go and for example collect soil samples, measure yields collect soil samples, measure yields collect soil samples, measure yields when you think about crop and farming when you think about crop and farming when you think about crop and farming and we're using this to calibrate a lot and we're using this to calibrate a lot and we're using this to calibrate a lot of what we see from satellites. of what we see from satellites. of what we see from satellites. The satellites that you're pulling this The satellites that you're pulling this The satellites that you're pulling this information from, are they just seeing information from, are they just seeing information from, are they just seeing the visible spectrum? Are they seeing the visible spectrum? Are they seeing the visible spectrum? Are they seeing information that might be stored but not information that might be stored but not information that might be stored but not being used? Are you squeezing like being used? Are you squeezing like being used? Are you squeezing like squeezing the fruit of these uh squeezing the fruit of these uh squeezing the fruit of these uh satellites for information that we satellites for information that we satellites for information that we didn't even know we had? didn't even know we had? didn't even know we had? Yeah, we are. So even if you look at the Yeah, we are. So even if you look at the Yeah, we are. So even if you look at the open source satellite from NAZA and the open source satellite from NAZA and the open source satellite from NAZA and the European space agencies, a lot of them European space agencies, a lot of them European space agencies, a lot of them are actually giving you a resolution of are actually giving you a resolution of are actually giving you a resolution of 10 meters. So what we're doing for 10 meters. So what we're doing for 10 meters. So what we're doing for example is that we're able to use um example is that we're able to use um example is that we're able to use um augmented techniques to do super augmented techniques to do super augmented techniques to do super resolution and look at um um and go down resolution and look at um um and go down resolution and look at um um and go down to a meter um resolution so we can see to a meter um resolution so we can see to a meter um resolution so we can see even more of what's happening on the even more of what's happening on the even more of what's happening on the ground. Um because one of the things to ground. Um because one of the things to ground. Um because one of the things to um uh understand is also the different um uh understand is also the different um uh understand is also the different context. So when you think about doing context. So when you think about doing context. So when you think about doing satellite analysis on the US for satellite analysis on the US for satellite analysis on the US for example, especially if you think about example, especially if you think about example, especially if you think about farming, it's large farms, monocrops. Um farming, it's large farms, monocrops. Um farming, it's large farms, monocrops. Um so it's quite easy to use open source

so it's quite easy to use open source so it's quite easy to use open source imagery and you don't have to um do a imagery and you don't have to um do a imagery and you don't have to um do a lot of uh fine-tuning with the image. lot of uh fine-tuning with the image. lot of uh fine-tuning with the image. But when you look at Africa, it's But when you look at Africa, it's But when you look at Africa, it's actually one by one square kilometer actually one by one square kilometer actually one by one square kilometer farms, multiple crops. It's not as farms, multiple crops. It's not as farms, multiple crops. It's not as structured and it's not as big. So it's structured and it's not as big. So it's structured and it's not as big. So it's very very difficult to go down the pixel very very difficult to go down the pixel very very difficult to go down the pixel level and really understand what is in a level and really understand what is in a level and really understand what is in a pixel. So my team has spent a lot of pixel. So my team has spent a lot of pixel. So my team has spent a lot of time building models that are actually time building models that are actually time building models that are actually taking this raw image um and um adding a taking this raw image um and um adding a taking this raw image um and um adding a lot of uh machine learning models that lot of uh machine learning models that lot of uh machine learning models that are allow us to understand things like are allow us to understand things like are allow us to understand things like land use land cover do things like crop land use land cover do things like crop land use land cover do things like crop masking being able to detect pest and masking being able to detect pest and masking being able to detect pest and disease like a lot of things that you disease like a lot of things that you disease like a lot of things that you would see are micro micro micro details would see are micro micro micro details would see are micro micro micro details on the on the environment we've been now on the on the environment we've been now on the on the environment we've been now able to combine both AI and data science able to combine both AI and data science able to combine both AI and data science with geospatial imagery to be able to um with geospatial imagery to be able to um with geospatial imagery to be able to um really understand and fine-tune what we really understand and fine-tune what we really understand and fine-tune what we can see from the sky. Now, I want to ask can see from the sky. Now, I want to ask can see from the sky. Now, I want to ask a dumb question and forgive me this a dumb question and forgive me this a dumb question and forgive me this might be a myth or a stereotype, but might be a myth or a stereotype, but might be a myth or a stereotype, but you're the perfect person to answer it.

you're the perfect person to answer it. you're the perfect person to answer it. I have heard uh as a person who lives in I have heard uh as a person who lives in I have heard uh as a person who lives in the northern hemisphere that like well, the northern hemisphere that like well, the northern hemisphere that like well, you know, there's cloud cover, there's you know, there's cloud cover, there's you know, there's cloud cover, there's there's jungles, there's things you there's jungles, there's things you there's jungles, there's things you can't see, like there's so much like can't see, like there's so much like can't see, like there's so much like it's impossible to see into the dark it's impossible to see into the dark it's impossible to see into the dark continent. I think that that's a continent. I think that that's a continent. I think that that's a narrative that we are told. Can you narrative that we are told. Can you narrative that we are told. Can you squeeze more information out of these squeeze more information out of these squeeze more information out of these satellites by multiple passes? Uh is satellites by multiple passes? Uh is satellites by multiple passes? Uh is there in fact cloud cover? Is there in there in fact cloud cover? Is there in there in fact cloud cover? Is there in fact places that you can't see into or fact places that you can't see into or fact places that you can't see into or do you have techniques that make that do you have techniques that make that do you have techniques that make that nonsense? We have techniques that make nonsense? We have techniques that make nonsense? We have techniques that make that nonsense. So it's true there's that nonsense. So it's true there's that nonsense. So it's true there's cloud cover. If you look at Nigeria and cloud cover. If you look at Nigeria and cloud cover. If you look at Nigeria and Brazil, there like those those two Brazil, there like those those two Brazil, there like those those two country actually are very similar in country actually are very similar in country actually are very similar in terms of uh vegetation. Um and they have terms of uh vegetation. Um and they have terms of uh vegetation. Um and they have a lot of cloud cover. So what we a lot of cloud cover. So what we a lot of cloud cover. So what we realized is that we had to do techniques realized is that we had to do techniques realized is that we had to do techniques such as cloud masking. We actually had such as cloud masking. We actually had such as cloud masking. We actually had to build our own version of Google Earth to build our own version of Google Earth to build our own version of Google Earth Engine because we needed to layer Engine because we needed to layer Engine because we needed to layer multiple different satellites to be able multiple different satellites to be able multiple different satellites to be able to get the insights that we needed. So to get the insights that we needed. So to get the insights that we needed. So our team spent so much time working on our team spent so much time working on our team spent so much time working on our geo package which is basically our our geo package which is basically our our geo package which is basically our geoprocessing engine that allows us to geoprocessing engine that allows us to geoprocessing engine that allows us to pick from all these different satellite.

pick from all these different satellite. pick from all these different satellite. So we're working with different So we're working with different So we're working with different satellite. We're working with um satellite. We're working with um satellite. We're working with um multisspectral as well. So you would multisspectral as well. So you would multisspectral as well. So you would look at like um satellite that are look at like um satellite that are look at like um satellite that are hyperspectral as well and being able to hyperspectral as well and being able to hyperspectral as well and being able to look at uh at satellite that are SAR um look at uh at satellite that are SAR um look at uh at satellite that are SAR um a synthetic aperture radar satellite. So a synthetic aperture radar satellite. So a synthetic aperture radar satellite. So being able to ingest those those being able to ingest those those being able to ingest those those different um resolution, different different um resolution, different different um resolution, different modalities of satellites, we've been modalities of satellites, we've been modalities of satellites, we've been able to really squeeze in the maximum we able to really squeeze in the maximum we able to really squeeze in the maximum we could from open source imagery um and um could from open source imagery um and um could from open source imagery um and um and run our analytics um um as we can. and run our analytics um um as we can. and run our analytics um um as we can. So this has been almost a year and a So this has been almost a year and a So this has been almost a year and a half in in the making. Our geo package half in in the making. Our geo package half in in the making. Our geo package now is fully running. That's what we're now is fully running. That's what we're now is fully running. That's what we're using for all our analytics in using for all our analytics in using for all our analytics in geospatial. But we also know that geospatial. But we also know that geospatial. But we also know that geospatial is not the only modality that geospatial is not the only modality that geospatial is not the only modality that you can use. So we've also spent some you can use. So we've also spent some you can use. So we've also spent some time working on ingesting drone imagery time working on ingesting drone imagery time working on ingesting drone imagery because you have a lot of drone because you have a lot of drone because you have a lot of drone companies not all across the continent companies not all across the continent companies not all across the continent but in some specific pockets uh being but in some specific pockets uh being but in some specific pockets uh being also able to ingest tabular data or even also able to ingest tabular data or even also able to ingest tabular data or even data from IoT sensors. A lot of people data from IoT sensors. A lot of people data from IoT sensors. A lot of people have deployed IoT sensors. You have have deployed IoT sensors. You have have deployed IoT sensors. You have startups who are doing um um uh electric startups who are doing um um uh electric startups who are doing um um uh electric bike charging and all their electric bike charging and all their electric bike charging and all their electric bike charger actually have a small bike charger actually have a small bike charger actually have a small weather station in there. So we able to weather station in there. So we able to weather station in there. So we able to capture a lot of that data because they capture a lot of that data because they capture a lot of that data because they have oh we have sensors that can measure have oh we have sensors that can measure have oh we have sensors that can measure rainfall temperature okay you have a rainfall temperature okay you have a rainfall temperature okay you have a small weather station can we use your small weather station can we use your small weather station can we use your data so this is the type of little hacks data so this is the type of little hacks data so this is the type of little hacks that we've been able to do and we've that we've been able to do and we've that we've been able to do and we've been extremely communitydriven been extremely communitydriven been extremely communitydriven um and building that community around us um and building that community around us um and building that community around us to help us really um strengthen our data

to help us really um strengthen our data to help us really um strengthen our data sets which then strengthens our outputs sets which then strengthens our outputs sets which then strengthens our outputs on on the machine learning side that on on the machine learning side that on on the machine learning side that that idea that there are all these that idea that there are all these that idea that there are all these electric bikes just sitting around there electric bikes just sitting around there electric bikes just sitting around there and they just happen to have a weather and they just happen to have a weather and they just happen to have a weather station. Makes total sense, but that station. Makes total sense, but that station. Makes total sense, but that scratches a part of my brain I didn't scratches a part of my brain I didn't scratches a part of my brain I didn't know that I had. That's such a clever know that I had. That's such a clever know that I had. That's such a clever clever hack. Yeah, it is. It is. You use clever hack. Yeah, it is. It is. You use clever hack. Yeah, it is. It is. You use this term cloud masking and this is like this term cloud masking and this is like this term cloud masking and this is like a well-known thing in the in the a well-known thing in the in the a well-known thing in the in the satellite uh you know uh data ingress satellite uh you know uh data ingress satellite uh you know uh data ingress kind of universe. The idea that you've kind of universe. The idea that you've kind of universe. The idea that you've got uh cloud shadows, you've got clouds got uh cloud shadows, you've got clouds got uh cloud shadows, you've got clouds themselves and then you want to themselves and then you want to themselves and then you want to eliminate them to ultimately get a eliminate them to ultimately get a eliminate them to ultimately get a cloud-free image, right? Is that the cloud-free image, right? Is that the cloud-free image, right? Is that the goal? A complete perfect cloud-free goal? A complete perfect cloud-free goal? A complete perfect cloud-free image of the entire continent down to image of the entire continent down to image of the entire continent down to the meter. So yeah, so cloud-free image the meter. So yeah, so cloud-free image the meter. So yeah, so cloud-free image not of the entire continent, but you not of the entire continent, but you not of the entire continent, but you would look especially so we don't when would look especially so we don't when would look especially so we don't when we start our analysis, we actually start we start our analysis, we actually start we start our analysis, we actually start at the micro level. So we will start at the micro level. So we will start at the micro level. So we will start with a specific country. Um we took a with a specific country. Um we took a with a specific country. Um we took a bit of a reverse approach because we bit of a reverse approach because we bit of a reverse approach because we realized that a lot of companies who are realized that a lot of companies who are realized that a lot of companies who are working with geospatial data actually working with geospatial data actually working with geospatial data actually start from the global approach. They start from the global approach. They start from the global approach. They start from the macro approach and that's start from the macro approach and that's start from the macro approach and that's how they train that's what that's what how they train that's what that's what how they train that's what that's what they use from a satellite imagery they use from a satellite imagery they use from a satellite imagery standpoint but also that's how they standpoint but also that's how they standpoint but also that's how they train their models. So we tested a lot train their models. So we tested a lot train their models. So we tested a lot of models that are global weather models of models that are global weather models of models that are global weather models global satellite imagery model global global satellite imagery model global global satellite imagery model global foundation models for earth science and foundation models for earth science and foundation models for earth science and what we realized is that when you're what we realized is that when you're what we realized is that when you're trying to look at downstream task in trying to look at downstream task in trying to look at downstream task in Africa the accuracy drops massively. So Africa the accuracy drops massively. So Africa the accuracy drops massively. So my team decided to take a very

my team decided to take a very my team decided to take a very contrarian approach and we said okay contrarian approach and we said okay contrarian approach and we said okay let's start small. So we'll start with let's start small. So we'll start with let's start small. So we'll start with one region in Kenya and then expand to one region in Kenya and then expand to one region in Kenya and then expand to the whole country of Kenya and then the whole country of Kenya and then the whole country of Kenya and then start expanding to other countries the start expanding to other countries the start expanding to other countries the region and then the the the continent. region and then the the the continent. region and then the the the continent. So we took a very uh um um opposite So we took a very uh um um opposite So we took a very uh um um opposite approach and we started understanding approach and we started understanding approach and we started understanding that there was a lot of micro details that there was a lot of micro details that there was a lot of micro details down to the pixel that we actually down to the pixel that we actually down to the pixel that we actually needed to understand if we wanted to needed to understand if we wanted to needed to understand if we wanted to have an analysis that was anywhere near have an analysis that was anywhere near have an analysis that was anywhere near like a good level of accuracy on the like a good level of accuracy on the like a good level of accuracy on the continent. Um so when we even started continent. Um so when we even started continent. Um so when we even started training our own foundation model we training our own foundation model we training our own foundation model we decide we realized that we needed to decide we realized that we needed to decide we realized that we needed to start with one country first and then start with one country first and then start with one country first and then expand to more country and fine-tune expand to more country and fine-tune expand to more country and fine-tune with a lot of ground truth data and even with a lot of ground truth data and even with a lot of ground truth data and even in in because a lot of countries in in in because a lot of countries in in in because a lot of countries in Africa don't even have the same climate Africa don't even have the same climate Africa don't even have the same climate the same vegetation like everything is the same vegetation like everything is the same vegetation like everything is very very different from east to west very very different from east to west very very different from east to west it's the complete polar opposite um east it's the complete polar opposite um east it's the complete polar opposite um east doesn't have much cloud masking west has doesn't have much cloud masking west has doesn't have much cloud masking west has quite a lot um so cloud cover sorry and quite a lot um so cloud cover sorry and quite a lot um so cloud cover sorry and west does have quite a lot. So we west does have quite a lot. So we west does have quite a lot. So we decided that we needed to really decided that we needed to really decided that we needed to really regionalize the way we approaching um regionalize the way we approaching um regionalize the way we approaching um our analysis and the way we're our analysis and the way we're our analysis and the way we're approaching building foundation model to approaching building foundation model to approaching building foundation model to just understand what's happening on the just understand what's happening on the just understand what's happening on the ground and then calibrate this with some ground and then calibrate this with some ground and then calibrate this with some of the global models that exist out of the global models that exist out of the global models that exist out there. Um and that has been a journey there. Um and that has been a journey there. Um and that has been a journey that our data science team has has that our data science team has has that our data science team has has undertook over the past two years. Like undertook over the past two years. Like undertook over the past two years. Like we didn't start and said we're going to

we didn't start and said we're going to we didn't start and said we're going to train a foundation model for science for train a foundation model for science for train a foundation model for science for Africa. We actually felt that we could Africa. We actually felt that we could Africa. We actually felt that we could use some of the ones that are available use some of the ones that are available use some of the ones that are available but we realized that the accuracy was but we realized that the accuracy was but we realized that the accuracy was not good. So that's why we decided to not good. So that's why we decided to not good. So that's why we decided to take that opposite approach and really take that opposite approach and really take that opposite approach and really do things like the continent is doing do things like the continent is doing do things like the continent is doing like developers of the continent are like developers of the continent are like developers of the continent are doing which is more bottom up than tops doing which is more bottom up than tops doing which is more bottom up than tops down and with this we've been able to down and with this we've been able to down and with this we've been able to reach some very decent and acceptable on reach some very decent and acceptable on reach some very decent and acceptable on our terms level of accuracy and we're our terms level of accuracy and we're our terms level of accuracy and we're now using this in our workflows. now using this in our workflows. now using this in our workflows. Now, you've mentioned open source a Now, you've mentioned open source a Now, you've mentioned open source a number of times and I've seen on GitHub number of times and I've seen on GitHub number of times and I've seen on GitHub like the Sentinel Hub cloud detector in like the Sentinel Hub cloud detector in like the Sentinel Hub cloud detector in Python and all kinds of things that I as Python and all kinds of things that I as Python and all kinds of things that I as a a person from the outside can look at a a person from the outside can look at a a person from the outside can look at can participate in. How much of the can participate in. How much of the can participate in. How much of the things that you're doing are proprietary things that you're doing are proprietary things that you're doing are proprietary foundation models that are private your foundation models that are private your foundation models that are private your your your own IP and how much of it are your your own IP and how much of it are your your own IP and how much of it are open source models that people can get open source models that people can get open source models that people can get involved in and and help. So, I'm a big involved in and and help. So, I'm a big involved in and and help. So, I'm a big proponent of open source. So I've spent proponent of open source. So I've spent proponent of open source. So I've spent a lot of my career working on on open a lot of my career working on on open a lot of my career working on on open source software, open source platforms source software, open source platforms source software, open source platforms as well. And um with our team, we as well. And um with our team, we as well. And um with our team, we decided that we would release our decided that we would release our decided that we would release our foundation model to the community. So foundation model to the community. So foundation model to the community. So it's actually been released last month it's actually been released last month it's actually been released last month on HuggingFace, the first version. Um on HuggingFace, the first version. Um on HuggingFace, the first version. Um now we're training now we're training now we're training V2. V0 is on Hugging Face. V1 is being V2. V0 is on Hugging Face. V1 is being V2. V0 is on Hugging Face. V1 is being trained right now. And we'll continue trained right now. And we'll continue trained right now. And we'll continue training and fine-tuning and releasing training and fine-tuning and releasing training and fine-tuning and releasing this to the community because we want this to the community because we want this to the community because we want people to contribute. Um we're not doing people to contribute. Um we're not doing people to contribute. Um we're not doing this for us. we're actually solving a this for us. we're actually solving a this for us. we're actually solving a challenge that we think we need more challenge that we think we need more challenge that we think we need more developers out there to support us to

developers out there to support us to developers out there to support us to do. Um and uh and that's why we actually do. Um and uh and that's why we actually do. Um and uh and that's why we actually um we made the conscious decision to um we made the conscious decision to um we made the conscious decision to release that with a research blog on our release that with a research blog on our release that with a research blog on our research and um our we were accepted research and um our we were accepted research and um our we were accepted this has been our first poster accepted this has been our first poster accepted this has been our first poster accepted to ICLR. So this in a two years company to ICLR. So this in a two years company to ICLR. So this in a two years company we're extremely proud to now starting to we're extremely proud to now starting to we're extremely proud to now starting to contribute to the research but also contribute to the research but also contribute to the research but also contribute to the open source community contribute to the open source community contribute to the open source community and we will continue doing so. That's and we will continue doing so. That's and we will continue doing so. That's amazing. So then how this is again a amazing. So then how this is again a amazing. So then how this is again a business type of question. You you business type of question. You you business type of question. You you you're not a nonprofit, right? You're you're not a nonprofit, right? You're you're not a nonprofit, right? You're not an NGO but you are committed to open not an NGO but you are committed to open not an NGO but you are committed to open source. How do you find that balance? source. How do you find that balance? source. How do you find that balance? So there is a balance between um the So there is a balance between um the So there is a balance between um the research that we're doing that we think research that we're doing that we think research that we're doing that we think can have an impact on all communities um can have an impact on all communities um can have an impact on all communities um including the local developer including the local developer including the local developer communities and there is a balance on communities and there is a balance on communities and there is a balance on serving our customers. So if you think serving our customers. So if you think serving our customers. So if you think about what our customers are buying from about what our customers are buying from about what our customers are buying from us, it's mostly insights. So you are us, it's mostly insights. So you are us, it's mostly insights. So you are global uh companies who exporting global uh companies who exporting global uh companies who exporting commodities whether it's agricultural or commodities whether it's agricultural or commodities whether it's agricultural or mineral out of Africa and you need more mineral out of Africa and you need more mineral out of Africa and you need more data on the first mile. So you would data on the first mile. So you would data on the first mile. So you would come to Amini for that. Or you're a come to Amini for that. Or you're a come to Amini for that. Or you're a government and you're trying to really government and you're trying to really government and you're trying to really change the way you're looking at change the way you're looking at change the way you're looking at infrastructure, you would also come to infrastructure, you would also come to infrastructure, you would also come to us because you want to understand like us because you want to understand like us because you want to understand like for the first time what are the actual for the first time what are the actual for the first time what are the actual resources that you have in your country.

resources that you have in your country. resources that you have in your country. Um so we really this is basically where Um so we really this is basically where Um so we really this is basically where we we we we are generating revenue is we we we we are generating revenue is we we we we are generating revenue is between global private sector companies between global private sector companies between global private sector companies and local government. The one thing we and local government. The one thing we and local government. The one thing we decided we will not do as a company is decided we will not do as a company is decided we will not do as a company is to make any of our communities pay. So to make any of our communities pay. So to make any of our communities pay. So developers working with us um are not developers working with us um are not developers working with us um are not paying. We're doing a lot of competition paying. We're doing a lot of competition paying. We're doing a lot of competition on Zindi. So it's one of the platforms. on Zindi. So it's one of the platforms. on Zindi. So it's one of the platforms. It's a bit like Kaggle, you know, um but It's a bit like Kaggle, you know, um but It's a bit like Kaggle, you know, um but very focused on on Africa and emerging very focused on on Africa and emerging very focused on on Africa and emerging economies. So we're working a lot with economies. So we're working a lot with economies. So we're working a lot with with Zindi to help train developers on with Zindi to help train developers on with Zindi to help train developers on geospatial data um and understanding how geospatial data um and understanding how geospatial data um and understanding how that specific modality works because that specific modality works because that specific modality works because it's not something that a lot of data it's not something that a lot of data it's not something that a lot of data scientists have touched in the past data scientists have touched in the past data scientists have touched in the past data scientists and machine learning scientists and machine learning scientists and machine learning engineers and it's much more complicated engineers and it's much more complicated engineers and it's much more complicated than computer vision. So we are trying than computer vision. So we are trying than computer vision. So we are trying to support our ecosystem to uh to uh to to support our ecosystem to uh to uh to to support our ecosystem to uh to uh to understand how to work with this understand how to work with this understand how to work with this specific type of data but also um specific type of data but also um specific type of data but also um supporting them with providing them some supporting them with providing them some supporting them with providing them some data sets so they can build models that data sets so they can build models that data sets so they can build models that are adapted to their countries and their are adapted to their countries and their are adapted to their countries and their communities. Africa is large and we're communities. Africa is large and we're communities. Africa is large and we're targeting not just Africa but also the targeting not just Africa but also the targeting not just Africa but also the global south. So now we're working in global south. So now we're working in global south. So now we're working in the Caribbean. We're starting work in the Caribbean. We're starting work in the Caribbean. We're starting work in Southeast Asia as well. So there is Southeast Asia as well. So there is Southeast Asia as well. So there is space for everyone and our our we have a space for everyone and our our we have a space for everyone and our our we have a big emphasis and we want to contribute big emphasis and we want to contribute big emphasis and we want to contribute to the community. So that's something to the community. So that's something to the community. So that's something we'll continue to do. Um so we've been we'll continue to do. Um so we've been we'll continue to do. Um so we've been finding that balance between the finding that balance between the finding that balance between the research the community but at the same research the community but at the same research the community but at the same time making sure we able to generate time making sure we able to generate time making sure we able to generate revenue and pay the bills. Very cool.

revenue and pay the bills. Very cool. revenue and pay the bills. Very cool. Hey friends, we're going to take a Hey friends, we're going to take a Hey friends, we're going to take a moment and hear from our new sponsor moment and hear from our new sponsor moment and hear from our new sponsor ShareGate at sharegate.com. I'm chatting ShareGate at sharegate.com. I'm chatting ShareGate at sharegate.com. I'm chatting with Shaylen Gimy. You know, uh, Jaylen, with Shaylen Gimy. You know, uh, Jaylen, with Shaylen Gimy. You know, uh, Jaylen, I've been thinking about moving my, uh, I've been thinking about moving my, uh, I've been thinking about moving my, uh, Google Suite over to M365, and, you Google Suite over to M365, and, you Google Suite over to M365, and, you know, the migrations, they're like know, the migrations, they're like know, the migrations, they're like pulling teeth. You know, what are the pulling teeth. You know, what are the pulling teeth. You know, what are the biggest headaches that IT teams will run biggest headaches that IT teams will run biggest headaches that IT teams will run into? And, and why does it take so long? into? And, and why does it take so long? into? And, and why does it take so long? You know, it's funny. I I find that IT You know, it's funny. I I find that IT You know, it's funny. I I find that IT admins always think migrations seem admins always think migrations seem admins always think migrations seem really complicated, but truthfully, if really complicated, but truthfully, if really complicated, but truthfully, if you do the planning really well, it can you do the planning really well, it can you do the planning really well, it can be smooth sailing. You know, it it's it be smooth sailing. You know, it it's it be smooth sailing. You know, it it's it all starts with planning. You need to all starts with planning. You need to all starts with planning. You need to figure out the scope, set realistic figure out the scope, set realistic figure out the scope, set realistic timelines, and take inventory of timelines, and take inventory of timelines, and take inventory of permissions. If you don't do that up permissions. If you don't do that up permissions. If you don't do that up front, you're kind of setting yourself front, you're kind of setting yourself front, you're kind of setting yourself up for delays and potential security up for delays and potential security up for delays and potential security risks down the line. Uh, you know, the risks down the line. Uh, you know, the risks down the line. Uh, you know, the reality is a lot of these migrations reality is a lot of these migrations reality is a lot of these migrations involve huge amounts of sensitive data. involve huge amounts of sensitive data. involve huge amounts of sensitive data. So, rushing isn't an option. And then So, rushing isn't an option. And then So, rushing isn't an option. And then there's also the Microsoft side. You there's also the Microsoft side. You there's also the Microsoft side. You know, Microsoft enforces some limits to know, Microsoft enforces some limits to know, Microsoft enforces some limits to manage traffic, which can slow things manage traffic, which can slow things manage traffic, which can slow things down even more. Uh that's where we come down even more. Uh that's where we come down even more. Uh that's where we come in. You know, we help IT teams get ahead in. You know, we help IT teams get ahead in. You know, we help IT teams get ahead of those roadblocks with uh ready-made of those roadblocks with uh ready-made of those roadblocks with uh ready-made reports like our site report lays out reports like our site report lays out reports like our site report lays out everything in your environment or uh everything in your environment or uh everything in your environment or uh source analysis which estimates how much source analysis which estimates how much source analysis which estimates how much effort your migration will actually effort your migration will actually effort your migration will actually take. Uh and to get around throttling take. Uh and to get around throttling take. Uh and to get around throttling well our customers can utilize well our customers can utilize well our customers can utilize concurrent migrations with ShareGate Pro concurrent migrations with ShareGate Pro concurrent migrations with ShareGate Pro allows them to accelerate their allows them to accelerate their allows them to accelerate their migration. Very cool. Well, that is migration. Very cool. Well, that is migration. Very cool. Well, that is ShareGate by WorkLE. It's one tool to ShareGate by WorkLE. It's one tool to ShareGate by WorkLE. It's one tool to migrate faster and secure your tenant.

migrate faster and secure your tenant. migrate faster and secure your tenant. It's an out-of-the-box Microsoft 365 It's an out-of-the-box Microsoft 365 It's an out-of-the-box Microsoft 365 migration and governance tool that you migration and governance tool that you migration and governance tool that you can learn all about at sharegate.com. can learn all about at sharegate.com. can learn all about at sharegate.com. Thanks a lot for being a sponsor. Now, Thanks a lot for being a sponsor. Now, Thanks a lot for being a sponsor. Now, you say that your custom AI models are you say that your custom AI models are you say that your custom AI models are tailored not just for the global south, tailored not just for the global south, tailored not just for the global south, but also for for industries. We've been but also for for industries. We've been but also for for industries. We've been talking about agriculture. you've got talking about agriculture. you've got talking about agriculture. you've got this model that is primarily focused on this model that is primarily focused on this model that is primarily focused on Africa, but as you start to go into Africa, but as you start to go into Africa, but as you start to go into Southeast Asia, as you start to do Southeast Asia, as you start to do Southeast Asia, as you start to do something like Brazil, are those models something like Brazil, are those models something like Brazil, are those models applicable because of the southern applicable because of the southern applicable because of the southern hemisphere or is the is the agricultural hemisphere or is the is the agricultural hemisphere or is the is the agricultural ecosystem so different that you're going ecosystem so different that you're going ecosystem so different that you're going to need a Southeast Asia model or a to need a Southeast Asia model or a to need a Southeast Asia model or a South Asia model that's different? So, South Asia model that's different? So, South Asia model that's different? So, they're actually applicable because of they're actually applicable because of they're actually applicable because of the um very similar climatic conditions the um very similar climatic conditions the um very similar climatic conditions and topographies and population of and topographies and population of and topographies and population of farming or mining communities. Um so farming or mining communities. Um so farming or mining communities. Um so when we started we only were focused on when we started we only were focused on when we started we only were focused on Africa but all of the sudden we get Africa but all of the sudden we get Africa but all of the sudden we get people asking us can you do Brazil, can people asking us can you do Brazil, can people asking us can you do Brazil, can you do Colombia, can you do the you do Colombia, can you do the you do Colombia, can you do the Caribbean, can you do Nepal and we're Caribbean, can you do Nepal and we're Caribbean, can you do Nepal and we're like well we can try and once we started like well we can try and once we started like well we can try and once we started digging into so one of the process that digging into so one of the process that digging into so one of the process that we have when we approach a country um or we have when we approach a country um or we have when we approach a country um or an area is that we start splitting it in an area is that we start splitting it in an area is that we start splitting it in agricological zones. So rather than agricological zones. So rather than agricological zones. So rather than splitting it in like um administrative splitting it in like um administrative splitting it in like um administrative levels, so countywide or district, we levels, so countywide or district, we levels, so countywide or district, we actually you have to split it in actually you have to split it in actually you have to split it in different zones that have the exact same different zones that have the exact same different zones that have the exact same agorological conditions. So when we agorological conditions. So when we agorological conditions. So when we started doing this for other countries, started doing this for other countries, started doing this for other countries, we realized that a lot of it if not 70% we realized that a lot of it if not 70% we realized that a lot of it if not 70% was extremely similar to what we had

was extremely similar to what we had was extremely similar to what we had already done in Africa. So what we already done in Africa. So what we already done in Africa. So what we needed was more ground truth to be able needed was more ground truth to be able needed was more ground truth to be able to fine-tune those models so we can to fine-tune those models so we can to fine-tune those models so we can expand the coverage and the use case. Um expand the coverage and the use case. Um expand the coverage and the use case. Um so that's why we spend a lot of time in so that's why we spend a lot of time in so that's why we spend a lot of time in building that foundation layer. So if building that foundation layer. So if building that foundation layer. So if you look at our foundation model can do you look at our foundation model can do you look at our foundation model can do things like land use land cover. Um we things like land use land cover. Um we things like land use land cover. Um we also looking into similarity search um also looking into similarity search um also looking into similarity search um using embedding. So a lot of it is using embedding. So a lot of it is using embedding. So a lot of it is applicable to different regions with applicable to different regions with applicable to different regions with similar topographies and similar similar topographies and similar similar topographies and similar climates. climates. climates. You know, uh, I'm thinking again about You know, uh, I'm thinking again about You know, uh, I'm thinking again about the bicycles with their weather the bicycles with their weather the bicycles with their weather stations, and I'm remembering a couple stations, and I'm remembering a couple stations, and I'm remembering a couple of years ago, my iPhone popped up and of years ago, my iPhone popped up and of years ago, my iPhone popped up and said that it had a barometric pressure said that it had a barometric pressure said that it had a barometric pressure system, and Apple was asking if it system, and Apple was asking if it system, and Apple was asking if it wanted access to it, presumably to fix wanted access to it, presumably to fix wanted access to it, presumably to fix their weather prediction, which makes me their weather prediction, which makes me their weather prediction, which makes me realize that there's a ton of Android realize that there's a ton of Android realize that there's a ton of Android phones running around on in the global phones running around on in the global phones running around on in the global south. That's that's all they're all south. That's that's all they're all south. That's that's all they're all data. They're all weather sensors. data. They're all weather sensors. data. They're all weather sensors. They're all but they're locked up. it's They're all but they're locked up. it's They're all but they're locked up. it's locked up behind the the device itself locked up behind the the device itself locked up behind the the device itself on the edge or if Google has it, they're on the edge or if Google has it, they're on the edge or if Google has it, they're not letting us know. Is that is that not letting us know. Is that is that not letting us know. Is that is that true? Are there little tiny weather true? Are there little tiny weather true? Are there little tiny weather stations walking all over the the global stations walking all over the the global stations walking all over the the global south? Depends the on the sensors they south? Depends the on the sensors they south? Depends the on the sensors they have and depends as well if they have have and depends as well if they have have and depends as well if they have connectivity, but that could be the connectivity, but that could be the connectivity, but that could be the case. Yes.

case. Yes. case. Yes. I'm just thinking there's a lot of I'm just thinking there's a lot of I'm just thinking there's a lot of information out there that is being that information out there that is being that information out there that is being that that is not being utilized to your to that is not being utilized to your to that is not being utilized to your to your point. Now, you're are you the only your point. Now, you're are you the only your point. Now, you're are you the only company that is doing this? Is it just company that is doing this? Is it just company that is doing this? Is it just that the northern hemisphere and Europe that the northern hemisphere and Europe that the northern hemisphere and Europe and America have not, you know, kind of and America have not, you know, kind of and America have not, you know, kind of turned Sauron's eye towards the south? turned Sauron's eye towards the south? turned Sauron's eye towards the south? Could someone swoop in and and and and Could someone swoop in and and and and Could someone swoop in and and and and um cause I mean trouble if uh they um cause I mean trouble if uh they um cause I mean trouble if uh they decided to think that this is a a decided to think that this is a a decided to think that this is a a problem that they need to solve. problem that they need to solve. problem that they need to solve. I mean uh right now we are the only I mean uh right now we are the only I mean uh right now we are the only company doing this, not on the company doing this, not on the company doing this, not on the agriculture side or the weather side. agriculture side or the weather side. agriculture side or the weather side. For us again we're data infrastructure For us again we're data infrastructure For us again we're data infrastructure companies. So environmental is the first companies. So environmental is the first companies. So environmental is the first type of data we decided to tackle. Then type of data we decided to tackle. Then type of data we decided to tackle. Then we'll start expanding. We've already we'll start expanding. We've already we'll start expanding. We've already started expanding to more than this. Um started expanding to more than this. Um started expanding to more than this. Um so for us it's really how can we go and so for us it's really how can we go and so for us it's really how can we go and close that data and compute gap on the close that data and compute gap on the close that data and compute gap on the continent and then give access to others continent and then give access to others continent and then give access to others to actually build solutions. I want to to actually build solutions. I want to to actually build solutions. I want to build an ecosystem. I don't want to build an ecosystem. I don't want to build an ecosystem. I don't want to build a product. I don't build I don't build a product. I don't build I don't build a product. I don't build I don't want to build 10 hundreds of models for want to build 10 hundreds of models for want to build 10 hundreds of models for different applications. I want others to different applications. I want others to different applications. I want others to do so. I want to be make sure that they do so. I want to be make sure that they do so. I want to be make sure that they have the right data sets that they have have the right data sets that they have have the right data sets that they have access potentially to the right level of access potentially to the right level of access potentially to the right level of compute to do so. So we'll give them the compute to do so. So we'll give them the compute to do so. So we'll give them the tool. We're bit a kitchen that builds tool. We're bit a kitchen that builds tool. We're bit a kitchen that builds the tools for everybody else and then the tools for everybody else and then the tools for everybody else and then they'll go and build. Um so when you they'll go and build. Um so when you they'll go and build. Um so when you think about that positioning definitely think about that positioning definitely think about that positioning definitely we are the only ones. And if you want a we are the only ones. And if you want a we are the only ones. And if you want a a comparison of a very similar company a comparison of a very similar company a comparison of a very similar company in a very different industry though um in a very different industry though um in a very different industry though um that has done the same thing you would that has done the same thing you would that has done the same thing you would look at for example a palunteer back

look at for example a palunteer back look at for example a palunteer back more than 10 years ago when they were more than 10 years ago when they were more than 10 years ago when they were sitting in government's rooms black sitting in government's rooms black sitting in government's rooms black rooms and digitizing data for the rooms and digitizing data for the rooms and digitizing data for the government and building things for for government and building things for for government and building things for for for the defense and the military for the defense and the military for the defense and the military industry. We're not in the defense and industry. We're not in the defense and industry. We're not in the defense and military industry definitely not. But military industry definitely not. But military industry definitely not. But Africa is where the US government was at Africa is where the US government was at Africa is where the US government was at that time. We still not capable of that time. We still not capable of that time. We still not capable of capitalizing on our data. We still don't capitalizing on our data. We still don't capitalizing on our data. We still don't really understand how to unlock the full really understand how to unlock the full really understand how to unlock the full value of our data. And we're all talking value of our data. And we're all talking value of our data. And we're all talking about leaprogging AI. Okay. Yeah. But about leaprogging AI. Okay. Yeah. But about leaprogging AI. Okay. Yeah. But there are steps to go through first. Um there are steps to go through first. Um there are steps to go through first. Um and we are trying to walk our ecosystem and we are trying to walk our ecosystem and we are trying to walk our ecosystem down this path to becoming AI ready. So down this path to becoming AI ready. So down this path to becoming AI ready. So then we can go and all start adopting then we can go and all start adopting then we can go and all start adopting cloud across all our workflows and cloud across all our workflows and cloud across all our workflows and across the entire continent but also across the entire continent but also across the entire continent but also reach a level of digitization where reach a level of digitization where reach a level of digitization where everybody can say okay Africa is truly everybody can say okay Africa is truly everybody can say okay Africa is truly liprog and now you have LLMs available liprog and now you have LLMs available liprog and now you have LLMs available in multiple African languages even the in multiple African languages even the in multiple African languages even the ones that are the least represented and ones that are the least represented and ones that are the least represented and such and there is an abundance of data such and there is an abundance of data such and there is an abundance of data that we've been able to capture and that we've been able to capture and that we've been able to capture and we've been able to really even the we've been able to really even the we've been able to really even the playing field for Africa where now it's playing field for Africa where now it's playing field for Africa where now it's easy to access weather data It's easy to easy to access weather data It's easy to easy to access weather data It's easy to even use Google maps. Let me tell you, even use Google maps. Let me tell you, even use Google maps. Let me tell you, it's it's still a challenge here. You it's it's still a challenge here. You it's it's still a challenge here. You have places where you get lost all the have places where you get lost all the have places where you get lost all the time, right? So, a lot of like these time, right? So, a lot of like these time, right? So, a lot of like these very simple things that we take for very simple things that we take for very simple things that we take for granted in the global north, we hope granted in the global north, we hope granted in the global north, we hope that by solving that data gap, um we we that by solving that data gap, um we we that by solving that data gap, um we we actually will be able to enable those

actually will be able to enable those actually will be able to enable those things to happen on the continent. I things to happen on the continent. I things to happen on the continent. I like that you said that you're building like that you said that you're building like that you said that you're building an ecosystem and not a product. And it's an ecosystem and not a product. And it's an ecosystem and not a product. And it's worth noting for folks that may not be worth noting for folks that may not be worth noting for folks that may not be familiar with your work that you ran the familiar with your work that you ran the familiar with your work that you ran the developer relations arm of Nvidia uh and developer relations arm of Nvidia uh and developer relations arm of Nvidia uh and focused on emerging areas. So you're a focused on emerging areas. So you're a focused on emerging areas. So you're a community builder and have been one for community builder and have been one for community builder and have been one for a long time. You worked at the Tiny ML a long time. You worked at the Tiny ML a long time. You worked at the Tiny ML Foundation. How do you are you doing Foundation. How do you are you doing Foundation. How do you are you doing like it's not really Amini's job, but like it's not really Amini's job, but like it's not really Amini's job, but I'm imagining like how do you go and I'm imagining like how do you go and I'm imagining like how do you go and build this data infrastructure company, build this data infrastructure company, build this data infrastructure company, but I'm I'm assuming if you still have but I'm I'm assuming if you still have but I'm I'm assuming if you still have that developer relations bone that you that developer relations bone that you that developer relations bone that you want to do hackathons and teach people want to do hackathons and teach people want to do hackathons and teach people GIS and like everyone's focused on GPTs GIS and like everyone's focused on GPTs GIS and like everyone's focused on GPTs and generating text, that's the hot and generating text, that's the hot and generating text, that's the hot thing right now, but you want to teach thing right now, but you want to teach thing right now, but you want to teach the young people how to how to work with the young people how to how to work with the young people how to how to work with this data. Are you are you doing this data. Are you are you doing this data. Are you are you doing anything like that? Are you working with anything like that? Are you working with anything like that? Are you working with the local communities? We are so that uh the local communities? We are so that uh the local communities? We are so that uh that programmer mentioned with Zindi, that programmer mentioned with Zindi, that programmer mentioned with Zindi, we're actually running hackathons and we're actually running hackathons and we're actually running hackathons and we're doing webinars. So I'm not the we're doing webinars. So I'm not the we're doing webinars. So I'm not the only devil in my team. Of course, I have only devil in my team. Of course, I have only devil in my team. Of course, I have a very big bias for de so a lot of my a very big bias for de so a lot of my a very big bias for de so a lot of my teams are actually community builders teams are actually community builders teams are actually community builders and that's what we love doing. So we and that's what we love doing. So we and that's what we love doing. So we spend a lot of time going like doing spend a lot of time going like doing spend a lot of time going like doing workshops. We work with a lot of workshops. We work with a lot of workshops. We work with a lot of research centers, universities where we research centers, universities where we research centers, universities where we go and present about the company and do go and present about the company and do go and present about the company and do workshops on geospatial and GIS. We do workshops on geospatial and GIS. We do workshops on geospatial and GIS. We do with Zindi. We've been doing a couple of with Zindi. We've been doing a couple of with Zindi. We've been doing a couple of different hackathons. The latest one different hackathons. The latest one different hackathons. The latest one that was launched is a crop detection that was launched is a crop detection that was launched is a crop detection challenge um for Kodiva. So challenge um for Kodiva. So challenge um for Kodiva. So differentiating for example Kakao uh differentiating for example Kakao uh differentiating for example Kakao uh from uh coconut tree. So being able to

from uh coconut tree. So being able to from uh coconut tree. So being able to really um uh teach those developers how really um uh teach those developers how really um uh teach those developers how to use geospatial data. Um so we've been to use geospatial data. Um so we've been to use geospatial data. Um so we've been still um doing a lot of dev and it's still um doing a lot of dev and it's still um doing a lot of dev and it's it's been part of our DNA as a company it's been part of our DNA as a company it's been part of our DNA as a company and I hope that as we grow we'll be able and I hope that as we grow we'll be able and I hope that as we grow we'll be able to do more of that. to do more of that. to do more of that. That's fantastic. I know that in the That's fantastic. I know that in the That's fantastic. I know that in the past in the past 10 years, everyone has past in the past 10 years, everyone has past in the past 10 years, everyone has been saying that we need more been saying that we need more been saying that we need more developers. We need more programmers. developers. We need more programmers. developers. We need more programmers. And uh lots of companies from the And uh lots of companies from the And uh lots of companies from the northern hemisphere have been running northern hemisphere have been running northern hemisphere have been running all over Africa trying to say we need to all over Africa trying to say we need to all over Africa trying to say we need to teach more people how to be developers, teach more people how to be developers, teach more people how to be developers, but we need more ML engineers, right? We but we need more ML engineers, right? We but we need more ML engineers, right? We need more machine learning experts on need more machine learning experts on need more machine learning experts on the continent that have the ability to the continent that have the ability to the continent that have the ability to to work with this kind of data. Yeah, we to work with this kind of data. Yeah, we to work with this kind of data. Yeah, we do. Um and if you think about one of the do. Um and if you think about one of the do. Um and if you think about one of the biggest wealths of the continent is biggest wealths of the continent is biggest wealths of the continent is actually our youth. It's the youngest actually our youth. It's the youngest actually our youth. It's the youngest continent in the world and it will continent in the world and it will continent in the world and it will continue to increase. So how do you continue to increase. So how do you continue to increase. So how do you actually make sure you provide them the actually make sure you provide them the actually make sure you provide them the right tools for them to be able to right tools for them to be able to right tools for them to be able to really um build a future that's really um build a future that's really um build a future that's different than what our parents and our different than what our parents and our different than what our parents and our grandparents have built up until now. Um grandparents have built up until now. Um grandparents have built up until now. Um and I truly believe that it's going to and I truly believe that it's going to and I truly believe that it's going to be via technology and I've already been be via technology and I've already been be via technology and I've already been doing so. They are so smart. A lot of doing so. They are so smart. A lot of doing so. They are so smart. A lot of them would learn machine learning and them would learn machine learning and them would learn machine learning and data science through nontraditional data science through nontraditional data science through nontraditional password institutions. They'll go to password institutions. They'll go to password institutions. They'll go to hackathon to boot camp. They'll watch hackathon to boot camp. They'll watch hackathon to boot camp. They'll watch YouTube doera classes you know um they YouTube doera classes you know um they YouTube doera classes you know um they would not go to school to learn computer would not go to school to learn computer would not go to school to learn computer science. They'll be self-taught data science. They'll be self-taught data science. They'll be self-taught data scientists and machine learning scientists and machine learning scientists and machine learning engineers. And for me that shows how engineers. And for me that shows how engineers. And for me that shows how hungry this youth is. Um hungry for new

hungry this youth is. Um hungry for new hungry this youth is. Um hungry for new things for challenges for access to things for challenges for access to things for challenges for access to tools and opportunities. And I hope tools and opportunities. And I hope tools and opportunities. And I hope we'll be able to be a catalyst for that. we'll be able to be a catalyst for that. we'll be able to be a catalyst for that. That's lovely. Um, this is an That's lovely. Um, this is an That's lovely. Um, this is an infrastructure company, but there are infrastructure company, but there are infrastructure company, but there are infrastructure challenges that can't be infrastructure challenges that can't be infrastructure challenges that can't be ignored on the continent depending on ignored on the continent depending on ignored on the continent depending on what countries that you're in. Now, what countries that you're in. Now, what countries that you're in. Now, you've picked Nairobi as your home base. you've picked Nairobi as your home base. you've picked Nairobi as your home base. Have you bumped into energy, Have you bumped into energy, Have you bumped into energy, sustainability, or infrastructure issues sustainability, or infrastructure issues sustainability, or infrastructure issues as you try to do this work? Because I'm as you try to do this work? Because I'm as you try to do this work? Because I'm presuming you want to do the work on the presuming you want to do the work on the presuming you want to do the work on the continent. You don't want to ship the continent. You don't want to ship the continent. You don't want to ship the data to Delaware and back. You prefer to data to Delaware and back. You prefer to data to Delaware and back. You prefer to do the work on compute that's in Africa. do the work on compute that's in Africa. do the work on compute that's in Africa. Yeah, we we do have a couple of data Yeah, we we do have a couple of data Yeah, we we do have a couple of data science workstations. like super science workstations. like super science workstations. like super computer farm here on our office. Um so computer farm here on our office. Um so computer farm here on our office. Um so we've had to be very very creative about we've had to be very very creative about we've had to be very very creative about how we powering that. Um we have a UPS how we powering that. Um we have a UPS how we powering that. Um we have a UPS and alternative power source because you and alternative power source because you and alternative power source because you still have um power challenges. So one still have um power challenges. So one still have um power challenges. So one of my hope for this podcast is that the of my hope for this podcast is that the of my hope for this podcast is that the power doesn't cut all of the sudden you power doesn't cut all of the sudden you power doesn't cut all of the sudden you know because it still happens especially know because it still happens especially know because it still happens especially it's been raining quite a lot the past it's been raining quite a lot the past it's been raining quite a lot the past uh the past few weeks. So there have uh the past few weeks. So there have uh the past few weeks. So there have been a few power cuts and on this once been a few power cuts and on this once been a few power cuts and on this once the power is gone then you have a the power is gone then you have a the power is gone then you have a generator but takes a couple of minutes generator but takes a couple of minutes generator but takes a couple of minutes to come up. Um so there's still deep to come up. Um so there's still deep to come up. Um so there's still deep challenges on the continent on top of challenges on the continent on top of challenges on the continent on top of power and electricity you would add uh power and electricity you would add uh power and electricity you would add uh connectivity there places um a lot of connectivity there places um a lot of connectivity there places um a lot of places uh across the continent places uh across the continent places uh across the continent especially in in in rural places where especially in in in rural places where especially in in in rural places where you still don't have access to internet you still don't have access to internet you still don't have access to internet right um so these are challenges we've right um so these are challenges we've right um so these are challenges we've had to work around um but we've been had to work around um but we've been had to work around um but we've been very hacky in a way to to make sure that

very hacky in a way to to make sure that very hacky in a way to to make sure that we're able to um to uh to power and and we're able to um to uh to power and and we're able to um to uh to power and and continue to develop a lot of our continue to develop a lot of our continue to develop a lot of our workflows workflows workflows Um, so now we have our supercomputers Um, so now we have our supercomputers Um, so now we have our supercomputers that are still running. Sometimes I come that are still running. Sometimes I come that are still running. Sometimes I come back to the office and I see don't touch back to the office and I see don't touch back to the office and I see don't touch foundation model training, you foundation model training, you foundation model training, you know. So, so yeah, it's uh it's one of know. So, so yeah, it's uh it's one of know. So, so yeah, it's uh it's one of the things we've had to work with, but the things we've had to work with, but the things we've had to work with, but also it's one of the things that gives also it's one of the things that gives also it's one of the things that gives us that little like very different us that little like very different us that little like very different mindset than companies in the global mindset than companies in the global mindset than companies in the global north in the northern hemisphere, you north in the northern hemisphere, you north in the northern hemisphere, you know, um because we know how to build know, um because we know how to build know, um because we know how to build with our local context in mind. And this with our local context in mind. And this with our local context in mind. And this is one of the things that you call tech is one of the things that you call tech is one of the things that you call tech for us. So if you think about a team for us. So if you think about a team for us. So if you think about a team that is well placed to really solve for that is well placed to really solve for that is well placed to really solve for that in Africa is not going to be a team that in Africa is not going to be a team that in Africa is not going to be a team that comes from outside. It's going to that comes from outside. It's going to that comes from outside. It's going to be team that was born from within. Yeah. be team that was born from within. Yeah. be team that was born from within. Yeah. Very scrappy, very focused, very hungry. Very scrappy, very focused, very hungry. Very scrappy, very focused, very hungry. Yeah. I love Exactly. So what is the, Yeah. I love Exactly. So what is the, Yeah. I love Exactly. So what is the, you know, they always talk about you know, they always talk about you know, they always talk about inventors having a Eureka moment where inventors having a Eureka moment where inventors having a Eureka moment where they say, "We did it. We did I'm curious they say, "We did it. We did I'm curious they say, "We did it. We did I'm curious in the last couple of years as you've in the last couple of years as you've in the last couple of years as you've been building these models, was there a been building these models, was there a been building these models, was there a moment that sticks in your head where moment that sticks in your head where moment that sticks in your head where you did something or the team did you did something or the team did you did something or the team did something and they called you, they something and they called you, they something and they called you, they said, "Come in, come in. We've cracked said, "Come in, come in. We've cracked said, "Come in, come in. We've cracked it." And then you got to see something it." And then you got to see something it." And then you got to see something or answer a question that was unanswered or answer a question that was unanswered or answer a question that was unanswered before.

before. before. So we have uh one of the first employees So we have uh one of the first employees So we have uh one of the first employees of Naza in the company. His uh like of Naza in the company. His uh like of Naza in the company. His uh like extreme, we call him the professor, he's extreme, we call him the professor, he's extreme, we call him the professor, he's extreme geospatial extreme geospatial extreme geospatial um engineer. Um and um he uh he was um engineer. Um and um he uh he was um engineer. Um and um he uh he was telling me about when I first started I telling me about when I first started I telling me about when I first started I was not a geopaceial expert background was not a geopaceial expert background was not a geopaceial expert background in data science and machine learning. So in data science and machine learning. So in data science and machine learning. So for me it was new for a lot of us it was for me it was new for a lot of us it was for me it was new for a lot of us it was new because it was a new type of new because it was a new type of new because it was a new type of modality we had to work with but he was modality we had to work with but he was modality we had to work with but he was telling me about all these challenges telling me about all these challenges telling me about all these challenges he's experienced with um the likes of he's experienced with um the likes of he's experienced with um the likes of like GE over the and other geospatial like GE over the and other geospatial like GE over the and other geospatial platforms over the past couple of years platforms over the past couple of years platforms over the past couple of years and he's like listen this thing is and he's like listen this thing is and he's like listen this thing is really hard to use the on boarding is really hard to use the on boarding is really hard to use the on boarding is really hard um there's a lot of it lacks really hard um there's a lot of it lacks really hard um there's a lot of it lacks flexibility it would be great if we can flexibility it would be great if we can flexibility it would be great if we can build our own and I told him I think we build our own and I told him I think we build our own and I told him I think we can and he said I don't know you know can and he said I don't know you know can and he said I don't know you know it's like we're not how do you go about it's like we're not how do you go about it's like we're not how do you go about building your own geo package geospatial building your own geo package geospatial building your own geo package geospatial processing engine said I'm pretty sure processing engine said I'm pretty sure processing engine said I'm pretty sure we can figure it out and um it was last we can figure it out and um it was last we can figure it out and um it was last year when I got a call and he he called year when I got a call and he he called year when I got a call and he he called and he said Kate we did it and for and he said Kate we did it and for and he said Kate we did it and for someone who had been in the field for so someone who had been in the field for so someone who had been in the field for so long knew so much and was just dreaming long knew so much and was just dreaming long knew so much and was just dreaming about something he could give to fellow about something he could give to fellow about something he could give to fellow like students researchers that was much like students researchers that was much like students researchers that was much more easy to use than all the tools out more easy to use than all the tools out more easy to use than all the tools out there um and can do the same if not even there um and can do the same if not even there um and can do the same if not even more than what those two do like coming more than what those two do like coming more than what those two do like coming to the realization that he was actually to the realization that he was actually to the realization that he was actually capable to build his own. This is that

capable to build his own. This is that capable to build his own. This is that for me was a really proud moment. It for me was a really proud moment. It for me was a really proud moment. It happened at this time. It happened when happened at this time. It happened when happened at this time. It happened when the data science team um launched their the data science team um launched their the data science team um launched their first foundation model, the first first foundation model, the first first foundation model, the first version of the foundation model. like version of the foundation model. like version of the foundation model. like those things that I call them the kids, those things that I call them the kids, those things that I call them the kids, my engineers. These kids didn't know my engineers. These kids didn't know my engineers. These kids didn't know they could actually achieve, you know, they could actually achieve, you know, they could actually achieve, you know, they're like this. We're usually they're like this. We're usually they're like this. We're usually consumers of technology. We usually use consumers of technology. We usually use consumers of technology. We usually use anything that's out there. We don't anything that's out there. We don't anything that's out there. We don't build our own. But just seeing how this build our own. But just seeing how this build our own. But just seeing how this helps shift their mindset um around helps shift their mindset um around helps shift their mindset um around actually I don't need someone to do it actually I don't need someone to do it actually I don't need someone to do it for me. I can do it by myself and I can for me. I can do it by myself and I can for me. I can do it by myself and I can do it even better and I can do it do it even better and I can do it do it even better and I can do it adapted for exactly what I need. um was adapted for exactly what I need. um was adapted for exactly what I need. um was a was a really proud moment. That's was a was a really proud moment. That's was a was a really proud moment. That's was my aha moment and I think I'm going to my aha moment and I think I'm going to my aha moment and I think I'm going to have more as we continue to like come up have more as we continue to like come up have more as we continue to like come up with new tools. with new tools. with new tools. It's been such a joy chatting with you. It's been such a joy chatting with you. It's been such a joy chatting with you. Your uh enthusiasm for the possibilities Your uh enthusiasm for the possibilities Your uh enthusiasm for the possibilities of the continent is absolutely of the continent is absolutely of the continent is absolutely infectious and I really appreciate you infectious and I really appreciate you infectious and I really appreciate you taking the time to chat with us today. taking the time to chat with us today. taking the time to chat with us today. Thank you for having me. That was great. Thank you for having me. That was great. Thank you for having me. That was great. We have been chatting with founder and We have been chatting with founder and We have been chatting with founder and CEO of Amini, uh, purpose-built data CEO of Amini, uh, purpose-built data CEO of Amini, uh, purpose-built data infrastructure for the global south, infrastructure for the global south, infrastructure for the global south, talking to Kate Ko, founder and CEO.

talking to Kate Ko, founder and CEO. talking to Kate Ko, founder and CEO. Absolutely fantastic work that they're Absolutely fantastic work that they're Absolutely fantastic work that they're doing. You can check them out at doing. You can check them out at doing. You can check them out at amini.ai amini.ai amini.ai ni.ai. And this has been another episode ni.ai. And this has been another episode ni.ai. And this has been another episode of Hansel Minutes in association with of Hansel Minutes in association with of Hansel Minutes in association with the ACM Bitecast. And we'll see you the ACM Bitecast. And we'll see you the ACM Bitecast. And we'll see you again next week. again next week. again next week. [Music]