Columbia University computer science professor and Data Science Institute affiliate Vishal Misra is an award-winning researcher, a serial entrepreneur, and a lifelong cricket fan and player. He has merged these three disparate areas of interest for a new venture—Ask Here First—which enables natural language queries of virtually any database.

By using natural language to interrogate structured data, Ask Here First’s AI-based approach has near-universal applications and could dramatically simplify search in a wide range of industries, including finance, media, marketing, and sports, according to Misra.

While his academic research focuses on computer networking, mathematical modeling, and performance evaluation, Misra’s entrepreneurial experience stems from his love of cricket and goes back to 1993, when he moved to the U.S. from India to study electrical engineering at the University of Massachusetts Amherst.

“Cricket is the most popular sport in India and one of the most popular sports in the world,” he said. “When I came to the U.S., I was starved for cricket updates and stats, but I could only get scores from print magazines.”

Misra and a few like-minded fans developed a system to post cricket scores online. Cricinfo, as their system was called, grew rapidly, was acquired by ESPN in 2007, and remains one of the network’s most popular single sport websites, bigger than its NFL and NBA portals.  

Despite Cricinfo’s success, Misra wasn’t satisfied. “Cricket is a statistics rich sport, but querying the Cricinfo database is still very complicated and time consuming. It doesn’t even have a mobile [user interface].”

Recent advances in machine learning, AI, and natural language processing led Misra to consider the possibility of using natural language to query the Cricinfo database. He reached out to ESPNcricinfo editor-in-chief Sambit Bal last summer to suggest a few updates.

Misra’s initial approach was to see if the newest natural language AI model, GPT3, could be used to query the database, but he found it to be inadequate. With some experimentation, he was able to develop a proprietary architecture to accomplish his goal.

“I soon realized that the architecture that I had was very general. It could go from natural language to any kind of structured query,” Misra recalled. “And then I knew I had a marketable product.” 

Misra has pulled together a team of Silicon Valley and Wall Street veterans with deep expertise in information technology and data science, including two Columbia computer science Ph.D. alumni, to implement the architecture he designed. In addition to transforming the Cricinfo interface, the team is also developing custom tools for other clients.  

Misra has found many parallels between being tenure-track faculty and being an entrepreneur. “Both involve working independently, and following your own vision,” he said. “As an entrepreneur and a professor, you may have funding support, but you’re not micromanaged. In both cases, you’re seen as the expert, and you need to deliver.” 

— Karina Alexanyan, Ph.D.