Galileo cofounders Yash Sheth, Atindriyo Sanyal, and Vikram Chatterji were working on different projects—various “models”—but they all faced the same problem.
Before founding Galileo, Sheth was at Google, working on speech-recognition models that recognize spoken words from audio containing human speech and transform them into text for more than 100 languages. Chatterji was working at Google AI on financial-services models that could analyze scanned Know Your Customer documents and parse out specific information. And Sanyal worked on Uber’s feature store, powering data quality improvements for thousands of models that would inform all of the company’s automated systems.
The problem? “Building a high-quality model that worked for enterprise solutions was a big pain for the teams,” Sheth says. “It was extremely manual and slow.”
In other words, as more and more teams begin integrating A.I. into their systems, they needed tools to discover and fix critical errors by analyzing the data from the model’s perspective, some of which can be the result of mislabeled data or “data drift,” which happens when data evolves in such a way that it renders A.I. systems less accurate.
Sheth and his cofounders faced this issue for more than a decade, but started thinking about solutions in 2020 when many teams had to scramble to fix their models suddenly as they became outdated following the unexpected event of a global pandemic. For instance, some models were trained to analyze pre-pandemic data, and they had to now be refreshed and retrained.
But that wasn’t easy to do. And that’s how the trio got the idea for their startup, Galileo, which has created a suite of products to provide an end-to-end platform for data science teams to apply generative A.I. and popular machine learning algorithms to their enterprise data.
“There are going to be times when people will have to refresh the models,” Sheth says. “And if it takes the companies months to do that, then A.I. is not going to be adopted across the world fast enough.”
The Fortune Founders Forum is a community of entrepreneurs chosen by Fortune’s editorial team to participate at the annual Brainstorm Tech conference, which took place in Deer Valley, Utah, in July. Our inaugural cohort was selected based on a variety of factors, including the potential impact of their companies, and reflected a diversity of geographies, sectors, and demographics.
