My name is Jyotsna and I come from an aeronautical aerospace background. My interest in tech startup entrepreneurship led me to do a master's degree at UC Berkeley in California, to understand advanced engineering and at the same time understand how to build these tech products into a company and follow it. That's when I discovered machine learning and I was fascinated to see what it could do.
I took another year just doing an advanced degree at Culpo Technique in France, to understand, learn and study machine learning and how it could be applied in many different scenarios.
After that, I was working with Airbus, using these machine learning applied mathematics approaches for autonomous flight.
It was around this time that the idea of SpaceSense came to me: it was in a conversation where someone talked about a certain scenario and we felt that it would be easy to do with satellite imagery. I also thought it shouldn't take me more than a weekend to do a proof of concept.
So I ended up doing a proof of concept and it was straightforward to build machine learning solutions, but I also discovered that a lot of the things which I assumed existed did not exist for satellite data and were 10 times more difficult in building a solution compared to using a normal image.
I tried to get a better understanding of whether people know that satellite data is there and that it can be so useful for the use cases they are talking about. So I spent a year trying to talk to a lot of different kinds of people from different industries like banking, retail, non-profits, agricultural companies, environmental companies and even construction companies, trying to understand the problems they face and do they know that they can use satellite data to solve those kinds of problems.
One story that stuck with me was that of a nonprofit called "GiveDirectly". One of the challenges they faced in the Africa Mission they were doing at the time was trying to provide aid to homes affected by flooding. The only way they knew where to go to help was through media reports.
For me it was completely outrageous because every time there's a flood in our country, the first thing that comes on the news is a map that shows with all red kinds of mapping showing where exactly the flood was and which regions got affected.
Normally in our cities, we would know which part of that city is heavily populated and where the housing is, and you would know where to go and support. And for me, the fact that this company couldn't have access to something like that was scandalous, and the whole reason was that the satellite data itself was not accessible.
That is why I feel that it is more and more important for something like Spacesense.ai to exist, where it builds the tools and necessary technology to make it simple and more accessible to be able to provide these kinds of solutions, which impact our everyday life.
This has become increasingly important with climate change, where everything changes so frequently that you can no longer depend on how things have always been done.
Access to technology like satellite data, especially when you combine it with machine learning capabilities, can open up many avenues in many ways in terms of how people can adapt, how companies can adapt and how organizations can help people adapt their day to day.
This is why I felt that there's a huge need for something that can be in the middle of a machine with technology and the understanding of the customer to build something that can bring this capability to everyone's hands.
And that's exactly what I'm trying to do with SpaceSense.
Practical examples of using SpaceSense
Today we are fully focused on enabling organizations and experts in the fields of agriculture, environmental monitoring, and infrastructure to use satellite data to build sustainable solutions that will help them better adapt to climate change.
To give you a few examples, a company that would an ideal customer for us is a digital agriculture company that would use SpaceSense and then combine it with their agricultural expertise to build a product that is either doing something like monitoring tens of thousands of crops at the same time and detecting any kind of anomalies at a huge scale. Or, in use cases more related to climate change, they use it to adapt the inputs that go into the fields, such as how much fertilizer should I use, when should I use it to minimize the impact on the soil and at the same time maximize production for the farmer.
As far as an environmental company is concerned, we are actively using SpaceSense tools to develop additional tools that address the issue of "how exactly can I examine the impact of a company or industry on the climate? How can I do that? Where is deforestation occurring, at what level is it occurring, and how can I partner with the government to help them use this intelligence on the ground?"
The same is the case with infrastructure companies, who are trying to understand how their mining or energy customers can reduce their impact on the environment so that they can continue to run the same operations: "where is the negative impact occurring and how can we reduce it?"
The way the interaction happens is that companies in these sectors use our tool: their developer is the one who uses our tool. The data scientists would use our tool to develop the final applications with their domain knowledge. The only thing we do to provide this accessibility is to eliminate the need for specialized expertise that would otherwise be required with satellite data.
With our tool, they don't need to hire any specialized expertise in satellite imagery: their existing teams and existing data scientist developers can use our tool to build solutions.
This is a huge pain point that we are reducing with these companies today.
SpaceSense partners and customers
In terms of our market, if you look at the terms, we are part of the earth observation downstream market, which means the market that conserves everything around applications of satellite imagery.
What is relevant for us is the submarket of tools and processing within this larger observation downstream market, which is estimated to be about 8.6 billion by 2030.
The kind of customers - companies - that we are targeting today are usually tech companies with highly innovative teams and creating digital products within their industry within agriculture, within environmental monitoring. For example, for ESG reporting carbon initiatives, climate-related solutions as well and also for infrastructure monitoring to adapt to climate change.
These are the companies that we are working with today and normally they have industry expertise. So in agriculture, they have agronomists who understand what their customer needs, and they have an understanding of what they can gain from satellite and they use our tool to first develop the product, tested on the field and see what works, what doesn't work and create the product. So that product development process and once that is done, they will also use space sense to then put this product in operations at a very wide scale.
For example in the case of the nonprofit company, they know exactly what they're looking for: they want to see where the houses are and match that with where the highest impact of the flood.
We give them the tools to very quickly get that information without needing a very specialized remote sensing satellite data expert.
We do quite a few collaborations across the industry where we do have a lot of synergies and with other data partners where we know that together with satellite data and AI, this additional data partner can add a lot of value.
For example, the weather data companies.
The different options available to companies for using satellite data
Today, if I am a company that is building a carbon credit ecosystem as a digital solution for carbon credits and I know that satellite data is very useful in completing the entire value chain in a more scalable way I have two main options.
The first possibility is that I create a whole team internally with expertise in satellite data, but also that will build the infrastructure to process this kind of satellite data on a large scale. Plus the mission learning engineers that will create the solution.
The second option is to say that this satellite data infrastructure is not my main core business: I'm going to outsource it. So I hired a consulting company that would do the same thing. They'll have the same team, but they'll charge me a little more than what would have cost me to build this internally.
So what of these options are very expensive today?
This is why a lot of the companies today in this space end up doing a choice three, which is not to use satellites at all and use what is possible, what is available either. We have empirical mathematical models that are not accurate: everyone knows that they are not accurate, but we use them.
Or the other option is where we don't use any kind of models, but we make certain assumptions, estimations or we don't even have any way to validate exactly how things are on the ground or exactly how much carbon credit am I getting. So it becomes unverified credits for example.
Or the other option is the more expensive one where we have personnel deployed on the ground that is going to every site in person and making those bringing that information on the tool.
Today is one of the ways that these companies create their solution and it's a good starting point, but this is not scalable for sure. They also understand this and that has become a bottleneck today.
That is why today, along with SpaceSense there are a few more companies are trying to make this solution more accessible and put this technology in the hands of these kinds of users that could benefit from it and give them another option where they can build what they need for their business without needing to spend that much money in building entire expertise in satellite data.
They could also outsource the expertise and infrastructure to a tool like SpaceSense. So apart from SpaceSense, few companies are trying to enable many more customers to be able to use satellite data for these kinds of solutions today.
Stakeholder response to SpaceSense
One of the reasons our solution is not as visible today is that it is relatively new. We see the problems of these companies and, in response to these problems, we are creating solutions like SpaceSense. These are tools, emerging ways that are not known by many people.
Terms of the feedback we're getting from our stakeholder is very positive and very encouraging because the main thing we keep hearing from them again, and again is saying: "Oh, this is so much better and because of this, we can develop a product very quickly and provide the solution to our customers quickly compared to it's taking a year or two years before we can have something scalable and usable with this kind of technology."
And I think that's very positive.
One of the things I'm very proud of, and it's also my main goal, would be to put, within two years, this kind of solution in the hands of all data scientists who are trying to build solutions for environmental monitoring and sustainable development, as opposed to today's people who don't use these tools at all.
So our goal is to say that instead of a hundred or a thousand satellite experts trying to do everything for every kind of use case, we want to give this application, give this capability to all data scientists developing climate-related solutions. So all climate scientists could use this technology to build a million different solutions to a million different problems. That is our goal and we are getting a very good initial response, both from the corporate users of our solution and from the individual data scientists using it saying, "Oh, now thanks to this tool I can do things that I couldn't do before without you."
So yes, it has been quite a positive response and we hope to be able to achieve our goals within two years.
The challenge SpaceSense faces
Right now our solution's not open for all of your users. So our immediate next step is to open our platform to all users by the end of summer so that by the end of 2025, we aim to provide our solution to over 2000 individual users: that's our goal.
At the same time, one of the main challenges we face in this regard is the knowledge of the users of these companies. They say that satellite data can be used exactly here or here because they have a basic high-level idea of "yes, satellite data could be interesting", but they need to also take the step of saying," let's try and see where this is interesting, can we try a few solutions?".
That step has to be taken and that is a major challenge for us to convince a lot of these companies to take the first step.
Our solution is being developed with this in mind because today the reason they don't take this step is that they feel this is complex and we are building it to simplify it.
They should feel comfortable and see that this is simple and It's not going to take too much time or effort.
That our biggest challenge to show all these companies that are taking the first step is not complicated at all.