One of Jane Street's most important teams is full of professional speed daters

Now that AI makes it easier than ever to structure messy data, sourcing useful alternative data is one of the most important roles in an electronic trading firm. Eric Mannes, who runs the alternative data team at Jane Street, recently appeared on the firm's podcast, Signals and Threads, to discuss the small, niche team which puts that data in the hands of traders.

💥Follow us on WhatsApp for news alerts.💥

Jane Street's alt data team is split into two broad categories: data engineering and data sourcing. Jobs in the latter camp are far more social: sourcers will go around speaking to the desks to understand what kind of data they need, then will attend various events and conferences where vendors pitch their datasets to them. For example, Mannes said that companies like Neudata, which catalogue the various available data sources, run "speed dating" events. Jane Street's data sourcers have "10, 15-minute dates in a row" where vendors explain their datasets, and the sourcers have to decide whether it's useful to any trading desks or whether it's unnecessary noise.

Alt data is a massive field brimming with noise. The most famous example of alt data is satellite imagery, but there are inherent issues and risks to using that kind of data. Mannes said that footage of Walmart parking lots, for example, is "really indirect. The thing you’re getting at is how many people showed up, which is different from how much they’re spending." Documents that a company files with the SEC are more structured and usable, but everyone has them, so you won't get much alpha on that front. The best data sourcers, therefore, should be able to find data that's usable, but obscure enough to give their traders an edge.

The field of usable data is expanding, particularly regarding text data, because of the improvement in large language models which can give structure to unstructured data. There are risks to using LLMs too; because of the massive amount of data they're trained on, there may be subtle biases, especially when backtesting. "The LLM knows what happened to Enron," Mannes said, and might incorporate retrospective knowledge instead of using the information that traders would have actually had at that time. 

CTO Ron Minsky confirmed that Jane Street has been building "super smart LLMs that are useful productivity tools," but has ways around that issue. It has a data "Superstore" which, among other things records all the information that Jane Street receives and the context in which it was received so that any retrospective changes to that data are not carried forward.

Jane Street's alt data team also includes a 20-person contingent of data engineers, who build the pipelines that process its data. Mannes said the team is looking for curious, investigative people; engineering skills are required but the "systems complexity" is not as high as it is in the firm's broader engineering organization. Mannes said people with "science or social science backgrounds, who do work with messy real-world data and try to make sense of it, are well suited to this." The team is also launching an internship track in 2027.

Have a confidential story, tip, or comment you’d like to share? Contact: WhatsApp: http://wa.me/442079977910 (+44 20 7997 7910), Telegram: @AlexMcMurray, Signal: @AlexMcMurrayEFC.88 Click here to fill in our anonymous form, or email editortips@efinancialcareers.com.

Bear with us if you leave a comment at the bottom of this article: comments are moderated intermittently by human beings. Sometimes these humans might be asleep, or away from their desks, so it may take a while for your comment to appear. You must take sole responsibility for comments you post on this site. We will take reasonable steps to weed out anything that we consider to be offensive or inappropriate.