When you evaluate a data provider, you’re really evaluating the people behind it. The research process, the product quality, and the transparency all stem from the team’s experience and judgment. That’s why we’re kicking off this series by looking at the career of ExtractAlpha’s founder and CEO, Vinesh Jha — because understanding where he comes from helps explain the approach you’re working with today.
From the early days of quant research
If you were active in quantitative investing in the early 2000s, you’ll remember the rise of StarMine. It was one of the first firms to take the messy world of analyst earnings forecasts and turn it into something measurable. Vinesh was StarMine’s Director of Quantitative Research, and he built many of the factor models that helped institutional investors separate genuinely skilled analysts from the crowd.
That experience shaped his view of what makes a signal worth paying attention to. A model needed more than an interesting narrative; it had to be rigorously tested, point-in-time, and economically intuitive. Those principles would follow him through later roles at Merrill Lynch and Morgan Stanley’s PDT Partners, where he designed systematic equity strategies and experienced firsthand the pressures of running money.
“Goal: Help investors uncover new sources of alpha in alternative datasets.”
Building ExtractAlpha
In 2013, he launched ExtractAlpha with a simple but ambitious goal: help investors uncover new sources of alpha in alternative datasets. At the time, there weren’t many independent firms doing this kind of work. A lot of providers from that era have since disappeared or been absorbed.
A decade later, ExtractAlpha is still here — which tells you something about the durability of the approach.
The team has grown steadily across Hong Kong, the US, Europe, and Canada, now including colleagues from both quant and fundamental backgrounds as well as our Estimize division. That mix is deliberate. It takes years to assemble a group that can both design cutting-edge models and explain them clearly to clients. The result is a team that focuses not just on research but also on making sure the outputs fit into your workflow — whether that means clean identifiers, careful timestamping, or accessible portfolio-level analytics.
“Most don’t survive the filters of history, intuition, and tradability. The ones that do become part of our product suite.”
Over the years, ExtractAlpha has evaluated hundreds of datasets. Most don’t survive the filters of history, intuition, and tradability. The ones that do become part of our product suite. Earnings-related signals remain a core strength, particularly TrueBeats®, which forecasts earnings and revenue surprises by identifying historically accurate analysts. Natural language processing models capture the tone and content of earnings calls and financial news, in English and Japanese. Regional signals, such as the Japan News Signal and Transcripts Model Japan, extend coverage into markets that remain less efficient and underpenetrated by quants.
For you as an investor, the value is in knowing whether your portfolio is implicitly exposed to these signals — or missing out on them. That’s why we publish live performance and regular retrospectives, so you can see how the models behave through different market regimes, including periods when crowded factors stumble.
A foundation built for the long term
The alternative data space is crowded today. But back in 2013, when ExtractAlpha started, the idea that data could be packaged into actionable signals for institutional investors was far from mainstream. A decade on, the firm’s continued independence is unusual. The reason is straightforward: the focus has always been on research rigor, transparency, and practical usefulness, rather than chasing hype or short-term growth.
Vinesh’s career has been about finding ways to measure what others miss, and building an organization capable of delivering those insights in a usable form. For clients, that means access to signals that have been tested, tracked, and proven across multiple regions and regimes. It also means a partner who has already done the hard work of cleaning and aligning messy datasets so they can be integrated quickly into a research or trading process.
This series is meant to highlight the people behind ExtractAlpha. Next time, we’ll turn to Kristen Gavazzi — who, like Vinesh, spent part of her career at StarMine, and went on to build a career at the intersection of product, sales, and global client engagement. Her story offers another perspective on how experience and curiosity come together to help investors find an edge.