[Techie Tuesday] How Prashant Warier of Qure.ai went from writing algorithms for logistics to solving healthcare issues using AI

This week’s Techie Tuesday features Prashant Warier, Co-founder of AI-based healthcare startup Qure.Ai. From working on pricing algorithms to healthcare AI, Prashant believes date is always the key.

For Prashant Warier, technology and science were just a part of everyday life. While there was no distinct ‘Eureka’ moment in his life, Prashant’s love for technology lies in its use. 

Today, he is the co-founder of qure.ai, a startup that uses Artificial Intelligence (AI) to make diagnostic imaging like X-ray and MRI affordable and accessible. 

“We wanted to apply AI to healthcare. More than that, I was keen on using deep learning, because it throws a lot of data to a neural network,” says Prashant, “The amount of data we generate on a daily basis can determine so many different things. To top it, algorithms in AI and deep learning have the ability to pick up patterns faster and evolve.”

Qure.AI has impacted more than 600,000 lives to date, and its technology is being used in over 20 countries for the screening of tuberculosis, triaging of trauma and stroke cases, and automated reporting of chest X-rays. 

Prashant Warier, Co-founder Qure.AI

The engineering way 

Born and brought up in Bhilai (now Chhattisgarh, then Madhya Pradesh), Prashant’s father was an engineer at a steel plant while his mother was a schoolteacher. 

While he cannot distinctly remember what he liked about it, science always interested him. A shy kid and a book lover, Prashant says he was good at mathematics, science, languages, and sports alike. 

He was in class 11 when he started working on computers. “We would work in the computer lab for one hour every week, and I learnt Pascal and BASIC. But, I didn’t own a computer until 2004,” says Prashant. 

In 1997, Prashant decided to get into IIT to pursue a bachelor’s in engineering. This was also the year that students had to give the IIT entrance exam twice as the first exam question paper was leaked. 

“The exams happened again in July. I don’t know how I did in the first attempt but I got through the second and joined IIT Delhi,” says Prashant. 

There, he took up the manufacturing science and engineering course, a mix of mechanical engineering and ops research. 

“In school, using a computer was more about games than programming. It was in IIT that I started focussing on computer architecture for my mathematics class. I would say I became reasonably good at it,” recalls Prashant. He adds that Applied Math was his strong suit. 

“I always aimed at going abroad as I wanted to study further and see what was there in the field of Applied Math and Computation. So, I applied at six universities and I got through Georgia Tech and the University of Southern California,” he adds. 

In 2001, he joined Georgia Tech as it was the best for industrial research and moved to the US. 

Prashant during Georgia Tech days

Working on logistics algorithms with math 

After finishing his master’s in only two semesters at Georgia Tech, Prashant decided to pursue a PhD. For this, he first worked on integer programming for large trucking networks in the US. 

He used several mathematical techniques to solve logistics issues for the trucking network in the US.

“When goods aren’t one full truckload, how to optimise shipments, and how are drivers assigned? There are several logistics rules as well. The drivers cannot be on duty more than 14 hours and cannot drive for more than 11 hours in the US, and there is huge trucking network that needs to optimise cost,” says Prashant. 

He explains that in those days, you would throw a problem at a computer, it would take longer. So, you needed to work on hacks with several mathematical combinations and techniques. 

“Programming for me is a way to solve a problem. I write code to solve problems. So, I began coding on a daily basis and started trying out new techniques,” recalls Prashant. 

Retail pricing for Esprit 

In 2007, he was going to join Kymetric, a price optimisation software company that would set up pricing for different retail stores in Southwest US. However, in the same year, the company was acquired by SAP and Prashant joined SAP to work on pricing for apparel companies instead. 

“I worked for Esprit. There are some apparel throughout the year while others are seasonal. We were trying to figure out manufacturing, SKUs, and how to price it. I see sales and depending on that, you either mark down or don’t give discounts,” he says. 

But by 2010, life had other plans. Prashant’s wife’s passport was stolen by an agent in Mumbai and sold to someone else. As a result, she couldn’t get back to the US even with a new passport. 

“I was in Arizona, trying to work with customs and homeland security, and it didn’t work. So, in 2011, I left SAP and came back to India,” says Prashant. 

Prashant and Pooja with the early team of Qure.AI

The world of Fractal Analytics 

Once back in India, Prashant joined Fractal Analytics, where he worked on solutions for the company’s retail vertical. Wanting to do something of his own, Prashant started Imagna Analytics in 2012, an adtech company that would go on to work with big ecommerce companies, using cookie data to figure customer behaviour patterns and target ads. 

“All the companies went mobile for a year or two. And mobile ad targeting at the time was difficult. We received an acquisition offer from Fractal, and I rejoined it as the Chief Data Scientist in 2015,” he says. 

But, the thirst to strike out on his own was still there and Prashant felt like he wasn’t playing to his strengths. Eventually, he decided to build a startup within Fractal. And that is how Qure was born. 

Qure progression 

In early 2016, Prashant began working on Qure. It was then he met Dr Pooja Rao who was working in Amsterdam, doing a lot of data science work in neuroscience.

“I was looking for people working in data science with a medical background. And in February 2016, we began the work. Technology is what makes us; we have about 20 data sciences and techies. Today, we have one of the largest immunology data in the world,” says Prashant.

In the beginning, Qure thought of different uses of AI and deep learning. “We thought of AI-powered toys for kids, a toy that would grow up with a kid, AI-based fashion imaging, and then we wanted to apply AI to healthcare,” says Prashant.

However, the team decided to focus on understanding images in healthcare in radiology and pathology. 

Today, while Prashant doesn’t need to code every day, he looks for people who have the ability to learn fast and have a learning mindset. “In this era, everything is easily available online. It is important to have someone who is adaptable and can learn faster,” he says. 

Saying that open-source is freely available, and anybody can read the latest papers, Prashant says, “Compared to 15 years ago, everything is easily available today. People need to adapt and learn quickly. So, learn and build expertise in one area, it helps.” 

Edited by Saheli Sen Gupta