← T-Minus: Space-Cyber Briefing
T-Minus: Space-Cyber Briefing
AWS in Orbit: Providing Decision Intelligence from Space with SatSure.
00:30:52
Find out how AWS for Aerospace and Satellite enhances SatSure's mission to provide decision intelligence from space.
Loading summary
Transcript31 lines
- [00:35]
Maria Varmazes
Welcome to AWS in Orbit. I'm Maria Varmazes. We're working with AWS to bring you an in depth look at the transformative intersection of cloud computing, space technologies and generative AI. On AWS in Orbit, we're exploring not just what's possible, but what's meaningful in the realm of space and cloud innovation. We grapple with the complex challenges and unparalleled opportunities that arise when we use space to address address pressing issues right here on Earth. This is AWS in Orbit providing decision intelligence from space with Sature.
- [01:28]
Rashmit Sukhmani
Hi Maria, My name is Rashmit Sukhmani. I'm a CTO and co founder of Satsur. Satchel is a deep tech company that provide decision intelligence from space. Our mission is to bridge information gaps and empower stakeholders to make informed decisions particularly in the challenging areas, especially like agriculture financing and climate risk management. So we started in 2017. Most of the customers are from big enterprises where we are leveraging Earth observation data to help them make wiser business decisions. What that basically means is we are helping them to solve their critical business problems, understanding their business, aligning Earth observation analytics to that and helping them to achieve a better top line or reduce their bottom line.
- [02:13]
Maria Varmazes
Paint me a picture of the challenges that your customers are facing and what are the problems they are looking to solve.
- [02:20]
Divya Sharma
My name is Divya Sharma and I currently head the Earth observation data team. We have customer in both verticals, agriculture as well as in energy and utilities. So the problem statements are very different and if you have seen this industry, they have evolved also very differently. Agriculture specifically in India, very traditional because of the small landholdings, because of the lack of mechanization, you can't do much about it. So one of the problems and it's very interesting and credit goes to the founders, how this particular problem was actually found and worked upon is our customer currently are banks. But the problem was actually found by interacting with farmers, specifically low income farmers. These low income farmers earn as low as $100 a month which you can imagine what all could you do. You have to run a family, you have to use it for inputs of your land mechanization if at possible. And all right. So the customer is still bank allowing them to get actionable inside to take a decision such as underwriting for a loan in scenarios where this low income farmer is not part of the credit system. So this is a very interesting sort of combination where the bank has a lot of money to disperse because there has been a mandate from through policies, through reserve backup India. But and farmers needs it. But then there is, there is no trust line between them that they it can be dispersed and you know, it can be trusted upon to give. So the bank is very blind to be able to tell whom should I give it to. So this is what we are trying to solve, trying to characterize the land parcel which this farmers hold. And since we're using earth observation data and decision insight, we have seen that land for so many years, so many seasons, we could really drive very actionable values, risk scores and stuff so that it can, you know, get the bank manager some more information.
- [04:23]
Maria Varmazes
Is there anything you wanted to add to that?
- [04:25]
Rashmit Sukhmani
We started a journey by understanding how agriculture is so complex, especially in the developing countries like India, right. And our whole idea was that how can we help the farmer? But then being a small team, you'll never be able to cater each and every farmer. Now what will be the entire alternate solution around it? How can you empower the whole system here to be much more agile, flexible and identify the right problems as such and then provide a solution around it? And that is why we started with BF as sector. But the intent was always the same, what Divya mentioned, right? How to help the marginal smallholder farmers and how to help them to have a better access to the credit system they may not have, let's say the bank accounts for that they do not have a credit score. But then now a credit manager are supposed to make that decision for them. So that was the idea and that's what our flagship products are also aligned to that.
- [05:25]
Maria Varmazes
That is such an important distinction. And I think when we talk about the way that data from space can inform these decisions, I think sometimes that chain of decision making is not always understood very well. But you both explained that so well and where that feeds in. So I really appreciate that, thank you. So let's talk a little bit about how these informed decisions are being made. How the data that you all are providing is being aggregated, for lack of a better term. Tell me a little bit about what that, what the tech stack looks like that is, you know, that is driving all these amazing decisions being made.
- [05:58]
Rashmit Sukhmani
It's not that simple. Okay. I think we have also learned that, right? The industry has been so conventional and traditional in terms of how to work with geospatial data. When I say geospatial, it has two facets, right? One, your remote sensing, the other one is a gis. But then nobody thinks about this data as more in terms of the product at the end of the day because you want to do that rinse and repeat kind of a work. Right. And that is where, that is where we started working at an intersection of geospatial domain, your AIML domain and engineering domain. And you are working with petabytes of data. That also means that you cannot work on your simple workstations to turn out that information, build complex AIML model to deliver that last mile connect. Right. That is where we have built the whole tech stack in a way where we consume these petabytes of data, build the models around it, do the last mile connect through let's say APIs, mobile applications or in a form of a report which helps and empowers the end consumer to use that analytics directly into their decision making framework.
- [07:11]
Maria Varmazes
That is, it is an incredible amount of data being moved and being data crunched. Divya, it looks like you wanted to add something to what Rashmit was just saying, so I wanted to give you that opportunity.
- [07:21]
Divya Sharma
Yeah, the problem statements which we are trying to solve, especially for the, for the countries as diverse as India, what really matter is how can you come up with more generalizable solution and hence the traditional technology is not going to make the cut. If it would have in the last five, 10 years, we would have seen really cool applications making a huge impact. So the time we are in, we are actually in a very, very sort of situationally good timing that AI ML has gone in production and with the right framework, with the right problem solving, you can actually and cross disciplinary teams, you can actually build very generalizable, very diversified solutions for the problems. So the nature of the problem demands.
- [08:08]
Maria Varmazes
That at this point and that would allow you, and I don't know if extrapolate is the right word here, but essentially to maybe even provide this decision intelligence in other settings that maybe you're not working in right now. Like in other geographies.
- [08:23]
Divya Sharma
Yeah, yeah, absolutely. So, so you don't start from scratch and it's easier to sort of fine tune or extend to different geographies.
- [08:31]
Rashmit Sukhmani
I just want to add this. See at the end people are paying us to give them the information in a way where they can consume. Right. Nobody will understand the maps or some numbers at the end of the day.
- [08:46]
Maria Varmazes
What does it look like for the end user?
- [08:47]
Rashmit Sukhmani
Yes, it's just a number so that they can use it in an actionable form. Right. And that is what we do even though we are building, but then a lot of massaging happens on top of it in terms of making sure that how the data is kind of related to one another, whether it Makes sense or not and deriving the decision. It's like a framework that you could think of and now it's a new buzzword of decision framework, decision intelligence, which has been recently coined. Data is the oil. But now moving forward it's an analytics which is an oil, which is powering a business.
- [09:21]
Maria Varmazes
That's a fascinating way of putting it. I like that metaphor. I like that. All right, so how does it we're talking petabytes of data, machine learning, allowing pulling new insights out of those petabytes of data in ways that we weren't able to do before scaling all of that. Also how does AWS help you all do all that?
- [09:41]
Divya Sharma
I think just as a building block itself. We've talked about the volume of data at which we are working, the scale of problem statements, the generalization, the diversity of the problem statement. We are working with petabytes of data and with these satellites always imaging Earth, especially the monitoring use case, they demand that kind of frequency. So the building blocks literally come from aws. I think without any of these cloud and compute provider and infrastructure, I don't think we'll be able to trying to aspire to solve at this continent level or a country level problem statement. Going back to aws, I think it has been really accelerating journey to be not worrying about how and what machines to use, where to deploy, how to structure my data and at the same time when we deliver the data to the customer, what should the be the inference or you know, the delivery pipelines looks like. So you're not worrying about the stability, reliability and scalability of the whole solution. You're more so in innovative zone and really problem solving. So I would say AWS compute, AWS storages and of course if I were to go deeper, more so inference, lambda pipelines and other things has really accelerated both our development journey as well as serving to our customers.
- [11:05]
Maria Varmazes
Fantastic. Thank you.
- [11:06]
Rashmit Sukhmani
And if I want to add a couple of more points on that, we're not just an Indian company where we are just serving Indian customers, whereas client is also global. Right. What that basically means is building a solution on an on prem infrastructure may not be the ideal way of doing things and working with the external stakeholders. Right. This is where AWS infrastructure helps us to go tap into different geographies, different markets and deploy the solution that would be one at the end of the day our USP is to work on the tech stack. When I say tech stack, build the pre processing pipeline, build the platform around it which can accelerate the journey of building the analytics AI ML Models. But what AWS brings on the table also is your security aspect, right. Which as a company like us may not have to worry about because at the end of the day we are working with a lot of financial institutions and different enterprises which take security very seriously. Right. And this is where the enabler like AWS helps us to do that.
- [12:16]
Divya Sharma
I think one of the product which has helped us really, really well is SageMaker AWS SageMaker which allowed us to do model training and model inference fairly quickly because there are pipelines built, they have already the container for the images of the model, model artifacts, all the access controls and the advanced configuration. So this allows us to do the model containers, model trainings as well as model inference in a very seamless manner. So it doesn't have to be like all the way orchestrated by us or configured by us. So it's literally, you know, it makes us life easier. Literally a button click thing. And then there is a serverless inference which allows us to save cost on our compute, so asynchronous inferences. So SageMaker has really nailed down the way to serve the data in the form and fashion we would like to. And in addition, we've been also toying around a lot of LLM models which has been housed on the SageMaker. If I'm giving data about India, you know, few states, 300 districts, nobody knows which district to pick up and which is performing bad or worse, right? So then we started really working on this very thin layer of LLM and to my surprise, it was just within two weeks we could find and actually develop, deploy a thin layer of decision intelligence which was based on AWS bedrock because there were already available some models which we just needed to prompt engineer or fine tune for our use case. And boom, there was a talking and chatting agent for our bank managers to be able to answer that query. So imagine the power of technology where we wanted to ideate on this conversational AI and two weeks after we have it. So this is the acceleration we were able to get through again due to AWS, SageMaker, AWS Bedrock and the services which AWS provides.
- [14:32]
Maria Varmazes
I would love to move into, if possible, examples of the impact that you all have seen on what you all are building on the lives of your customers. I don't know how specific you can get, but just any kind of customer example would be really great.
- [14:47]
Rashmit Sukhmani
See, we started our journey with bank at the end of the day, the reason being coming from the geospatial industry, understanding how the banking ecosystem worked. This is where we understood there There are gaps in the whole process. Right. And those gaps are not because they're not doing their job, they are doing their job. It's just that there's a lack of alternate data which empowers them to make a wiser decision at the end of the day. And this is where understanding the processes, identifying like especially what when TV touched upon underwriting, monitoring and better collection, these are the critical gaps because at the end of the day, what are you doing for the financial institution, you're saving them the money so that they can use that same money to empower the farmers with a better credits. Right. With better assessment. They don't make any wrong decisions at the end of the day. So this is where the impact and this whole journey we started working with some fewer banks and working with let's say 10, 15 districts right now we are working throughout the India. But in that whole thing we were roughly able to touch upon the lives of 20 million farmers as such, which, which were helped through the solution that we were providing. That's one part of that. And second would be if I take an example of some other sectors as such, for example utility, same business model, the problems are different, it's complexities just because how they're consuming the data, what sort of data. So at the end we were able to reduce the operational costs, help them to achieve the better top line. In a nutshell. See at the end, dollar is saved or dollar is grown, that's what the impact could be, I would say assessed as well as definitely what you're doing for the society. And this is what we have been doing, I think.
- [16:40]
Divya Sharma
Yeah. The biggest impact on scale we have seen, as Rashmi was mentioning, is by impacting farmers life by monitoring these kind of land parcels. And just to sort of give you a scale, we actually monitor about 85 million hectares, you know, every season. And he was mentioning the number of farmers. The hope is to support number of farmers as big as Germany's population. Right. So we're hoping to go. India is home to 120,130 million farmers and we are hoping to get to at least 80 to, you know, to really empower the farmer for better likelihood and more sustainable practices in agriculture. So yeah, because the thing has been felt there, the other things which I know we haven't talked about here, but we've been closely looking at energy and utility sectors where they need monitoring for vegetation management. So a lot of vegetation overgrowth around the conductors, around the high tension lines, distribution lines. And so these things can be easily sort of Monitored at the scale of the entire country. So we do have clients for, for whom we are actually monitoring these things over the entire country scale. So the impact is again, saving those billions of dollars, which are either the property loss, people loss, home, the land getting rendered useless because you can't grow anything. Right. So the other side then, the agriculture and the other side is also at the scale of the country is what we are talking about.
- [18:15]
Maria Varmazes
A word that has come up a lot, and I know it comes up a lot on like Satcher's website also is sustainability, which is a word that is used a lot in the space sector. And people have slightly different definitions of what it means to them and what they are doing. I would love to hear from you both what it means to you. So, Rashmit, I'd love to hear your thoughts on that.
- [18:35]
Rashmit Sukhmani
For us, sustainability is more in terms of how you're doing things, how reliable you are, and how empowering at the end of the day to your customers. And definitely those seven sustainability goals are kind of at the center here. If I talk about food security, if I talk about climate change and the impact it creates, this is where the solution that we are building takes all this into consideration in terms of understanding what has happened in the past, what impact now the climate could create in terms of your soil moisture reduction or your production of the crops, or a number of other things. Right. And this is where the whole definition of sustainability comes into picture for that show.
- [19:32]
Divya Sharma
Yeah. So I think when I have looked at sustainability, it's very different dimensions, very orthogonal. One dimension is how do you build your company, your solutions, so that you are not putting a lot of, what do you say, load onto the entire space chain, space and satellite chain. What I mean by that, the company has started with a notion of, first let's try to understand the problem. And without naming the companies across the world, you have seen the other way around. Everybody's so excited to launch a satellite. Everybody's so excited to take a picture of globe in night and day and publish. But at the end of day and Today we have 300 plus odd optical and maybe 25,000, 26,000 altogether. There's a mess up there. There's debris everywhere. Right. They're colliding. And nobody is using the whole idea of, you know, who's responsible for that. So I think the idea where Satya came in is, okay, we know we will have to at some point be self sustainable to get the data from space, but let me be more sustainable in the downstream first. So are there Enough problem statement or are there enough scalable solutions we can provide to these very specific problem statements which Rashmit has touched upon? Agri lending, angry financing, vegetation management, feature detections for urbanization, urban planning and all of that. So that was our word for sustainability. Let's work out the scale and volume and problem statement first, then for business continuity let go up in the space. And that is why we have our sister company called Kaleidio through which we are launching our own satellite, because we have seen sentinels going down and of course we can't be in a business where we can't control the entire chain. Hence we are wanting to be a full scale or full stack earth observation company. Now the other sustainability which I was talking about comes from the domain. So we are in agriculture domain and fortunately enough since we have been seeing this data for many seasons, many years, so we do see the drift in the crop performance by a mere 1 degree Celsius change. So now what can we again give back to the banks and other companies, insurance and other companies to really get that actionable understanding of what a 1 degree Celsius has done to a crop which has been growing for last 6 years? Has the yield reduced? Has there been more water demand? All of that is what we are trying to understand. So we didn't start with the very upfront notion of let's work on climate change. It just so happened that the data is telling us on our face. I'm changing every year, my yield is impacted, I'm not able to grow as much or my quality is bad, disease is coming. So this is another way of incorporating sort of climate information to give these insights to say, okay, this is the worst performed area. You have to pull in money, you have to invest in here so that the whole district or subdictors can come up with a more sustainable or more sort of bounce back to the livelihood they would. So you need to reinsure more or give more insurance. So yeah, they're very, very different thoughts. But yeah, I mean for us these two really, really matter as we develop and grow things.
- [22:55]
Maria Varmazes
Fantastic. Well, I think sort of naturally we've landed at the part where we talk about sort of long term plans and divi, you mentioned it a little bit in terms of ongoing vertical integration. So let's talk about, let's maybe start with long term goals. As a company you've already sort of hinted at some of that and then maybe long term goals for the real world impacts that you want sature to have. So that vision. So Rashmit, I imagine you might want to start with that one.
- [23:23]
Rashmit Sukhmani
See definitely we want to become a full stack company but by building let's say Earth Observation data refinery ecosystem which is designed to deliver the analytics ready data products that offers contextual, industry specific solution. Let me deep dive more into that. Right, sure. Divya mentioned about that vertical integration. Right. There's a reason why we are doing it in terms of controlling the upstream side as well so that we can do it better, so that we can tap into various complex business problems. Yes, we started with agriculture, but that was the first problem that we could see in front of us. But similar problems are there in different verticals as well and different sectors as well. For example, you can talk about infrastructure, utility management or forest management, that's another thing. Right at the end it's still aligned towards the sustainability goal. So we want to become that at the end of the day where we are in a control and building a tech stack which is so robust that it helps any sort of business where there's so much of complexity and making sure that at the end of the day geospatial data is not just for the consultation, it can be actually used in your day to day activity. And I think the world is also moving there because of the technology that is evolving around it. People talk about big data, aiml, the kind of innovation that is happening. The computing costs are going down. Right. So we want to leverage that and do it at scale solving various problems and helping the society. At the end of the day, our.
- [25:05]
Divya Sharma
Personal vision is to democratize the decision intelligence. The data sitting in analysis is no good to anyone. And specifically with the amount of watching we are doing of Earth, there's lot of impactful and disruptive and innovative technology we can deploy which can bring a huge scale impact. And we have seen it personally. So if we were to ask like personally. My vision for my team being the head of the composition data team is to really be able to come up with platform. While we are learning the solutioning of certain very difficult, very traditional industries. Anybody should be able to work off this platform and be able to solutionize for the problem statements. You know they have seen, I've seen problem statements such as tracking the spread of diseases that is being done long time using geospatial and who would have thought from medical to healthcare to agriculture to energy utilities, the potential is just immense. It's like sky is the limit at this point. And what we want to be is the enabler of the tool sets, the platform, the data. Since we are also trying to be a full stack company, be able to give that power and that democracy to be able to solution for the new problem statement. I think the problem statement and the, you know, the volume of the game is so much that it's not gonna take one person to build that. It's gonna take a village to build the solutions around. So that's why we happy to collaborate as well and we're looking forward to more such, you know, companies like AWS to accelerate that journey so that we work on what we do best as opposed to worrying about how to scale, how to sustain it and how to make it reliable systems.
- [26:48]
Rashmit Sukhmani
Yes, we have been talking about this industry being very traditional as such. But one thing in this whole journey I've learned that your technology can be swapped. What that basically means is the See all this analysis of an imagery of a remote sensing GIS came from the medical imaging at the end of the day, right. And innovations are happening there as well. So I would say that to an audience that even though it's still remote sensing and gis, but lot of different components of the technology which has been used in different sectors can be applicable here. Right. And this is where how AWS is helping us because they are bringing those partners to us, opening up the ecosystem which jointly built this whole community of geospatial space by integrating different components of the technology. Whether it's your medical imaging software engineering. Nobody would have thought about it 10 years back, right? People may would have been doing it which was like a military intelligence. But yes, now things are getting bit more commercial, people want to use it. So that is where it is essential that having an open mind that technology can be merged and can be utilized in different aspect altogether.
- [28:13]
Maria Varmazes
And that's it for AWS in Orbit providing decision intelligence from space with Sature. A special thanks to Divya Sharma and Rashmit Singh Sukhmani for joining us today. For additional resources from this episode and for more episodes in the AWS in Orbit series, check out our show notes@space.n2k.com AWS this episode was produced by Alice Carruth and powered by aws. Our AWS producer is Lara Barber. Our associate producer is Liz Stokes. We're mixed by Elliot Peltzman and Trey Hester with original music by Elliot Peltzman. Our executive producer is Jennifer Iban. Our executive editor is Brandon Karpf. Simone Petrella is our president. Peter Kilpie is our publisher and I'm your host, Maria Varmazas. Thanks for listening. We will see you next time.