That's the business side of their story. But the Wadhwani brothers have also carved a niche for themselves in philanthropy, giving back to society by capitalising on their core strength -- technology. Their not-for-profit start-up, the Wadhwani Institute of Artificial Intelligence (WIAI), which was inaugurated by Prime Minister Narendra Modi last year, aims to leverage the power of AI to benefit the social sector.
The Institute, which recently won a $2 million grant from Google.org, is now looking to expand into Africa based on it learnings in India in the areas of improving crop yields, mother and child health in rural and semi-urban areas, and tuberculosis. Headquartered in Mumbai, WIAI identifies important use cases in agriculture, education and health, among other fields.
“Everything we do is based on the understanding that it will get scaled and deployed,” says CEO Padmanabhan Anandan.
Set up a year and a half ago with an initial $3 million funding from the non-resident Indian brothers, the start-up has managed to build impressive partnerships in India and elsewhere. For example, it has an ongoing project with the Maharashtra government’s agriculture department, Welspun and other partners to help cotton farmers in the state. Another project looks at low birth-weight babies while the organization has recently collaborated with the Union Ministry of Health to make tuberculosis systems AI-ready.
WIAI hopes its projects, once completed, will be able to help the Maharashtra agriculture department decide when to issue an advisory to farmers about using pesticides in cotton fields. Similarly, through its collaboration with the Union Health Ministry, it is looking at helping the local governments record the accurate weight of babies through identical 3D modeling of newborns. In tuberculosis, the aim is to help the Central Tuberculosis Division predict which TB patients are likely to not adhere to their entire medicine course.
WIAI is close to running some AI field trials of its work in the cotton farms, while it has got the go-ahead for local health workers to shoot videos and pictures of babies in rural areas. This will help create 3D models of the babies that will be useful in accurately measuring their birth weights. The problem is an interesting one, given that many babies don't get measured at all. The other issue is that the weight of a newborn in these areas is measured on a spring balance, which is not only less accurate but risky too.
“Our goal is not to limit ourselves to India. By next year, we may be able to export some of the solutions to Africa and other countries. First we have to prove something...Once you get the basic framework, the model can be tweaked for another community,” Anandan said.
In addition to the initial funding by the Wadhwani brothers and the $2 million grant from Google AI Impact Challenge, WIAI is also approaching USAID, the World Bank and the Bill and Melinda Gates Foundation.
Anandan says, he is confident of getting enough funding, given that very few start-ups work in the social sector, but a bigger worry is the lack of people skilled in AI. “My bigger challenge is finding the talent. That’s why, I need good partnerships. AI talent is very hard to come by,” he says.
Given that it works with a small team of 32, WIAI has been practical about how it invests its available resources.
While the software or apps that they make to solve the problems are built by them, Anandan says having a reliable cloud provider is just as essential.
"We are a startup, you will rely on something like Amazon Web Services as the cloud where you do your building. You’re not going to build things yourself. It’s something we can configure quickly, and when you’re dealing with sensitive private data like health data then you need security and privacy and they have that. If you’re dealing with government data then govt needs to allow us to do it and AWS is empaneled with Ministry of Electronics and Information technology so that helps,” he says.
Its projects will require large data sets, but most data in the rural or agriculture sector is not easy to come by.
In Anandan’s words, it is either sparse and incomplete, unreliable, noisy, in silos, in heterogenous systems, not easy to get to, especially in health. He adds that there is sometimes no data at all.
However, the CEO is optimistic. “Our goal is not to limit ourselves to India. By next year we may be able to export some of the solutions to Africa and other countries. First we have to prove something...Once you get the basic framework working, We can retrain that model and get it trained for another community,” Anandan said.