Building a Global AI Venture for Medical Imaging Analysis from India: Prashant Warier, CEO of Qure.ai (Part 2)

Sramana Mitra: What happened to the company?

Prashant Warier: I got an exit in 2015. Somebody wanted to acquire the technology and use that for customer targeting for email engines.

Sramana Mitra: Was that bootstrapped?

Prashant Warier: Yes.

Sramana Mitra: You were doing it out of Mumbai?

Prashant Warier: Yes.

Sramana Mitra: Did you get a good exit?

Prashant Warier: It was okay. It was definitely some value.

Sramana Mitra: The exit happened in 2015. Did you start Qure in 2016?

Prashant Warier: The exit happened in 2015 to a company called Fractal. Fractal wanted me to be their Chief Data Scientist. I was excited. That was a good role. When I got into Fractal, I felt like there is a better opportunity for a data scientist to build products. Fractal was a services company. I was chatting with the CEO of Fractal about how we could build more products in that company. That is how Qure came about. The Fractal CEO and the investors wanted to incubate it. That was how Qure got its start.

Sramana Mitra: What was the premise of Qure?

Prashant Warier: I was working on image AI even in my previous company. If you look at Vogue, there is a model wearing a dress. Can I identify that and link it to an actual product on an e-commerce site? I wanted to see if we could take this into healthcare. We started with radiology. Can we apply the techniques to radiology? Can we train the algorithms to interpret results? That was the hypothesis.

When we started, we were working on radiology and pathology. Over time, we realized that the commercial paths are very different for both. We decided to stick to radiology. Any hospital system has historical radiology images. Radiology workflow is already digitized. From the time a scan is taken to the time it goes to the patient, that whole workflow is digitized.

Sramana Mitra: In terms of getting the dataset to train your model, how did you get it? Did you start working with a handful of customers who gave you the dataset? In all AI companies, one of the first questions that comes is, where is the data?

Prashant Warier: We were fortunate. We started out in India. In India, the magnitude of patients that come to hospitals is more than most other geographies. Every hospital group in India has got massive amounts of data. Radiology data is something that they always store. We had access to all these data.

It was comparatively cheaper back then because there was no real value to it. Now when you try to get data, it will be super hard. Back in 2016, there was no market for this data. We partnered with hospital groups and we were able to get good-quality data.

Sramana Mitra: Did you have to buy the data?

Prashant Warier: Yes. Some of it was revenue share. Any product that we’ll build out of it, we’ll get a percentage of revenue. All of it is anonymized. We don’t have patient identifiers. We don’t even have the patient history. We only have the scan and the corresponding report.

This segment is part 2 in the series : Building a Global AI Venture for Medical Imaging Analysis from India: Prashant Warier, CEO of Qure.ai
1 2 3 4 5 6 7