Jane Street CTO: "Everyone knows some Python... lots of people aren't any good at it."
Jane Street's continual evolution is not only happening on the trading floor. The market making firm has made significant changes to its engineering culture in recent years by introducing Python, which has been very useful for the firm's machine learning efforts, but presents a few other problems.
💥Follow us on WhatsApp for news alerts.💥
"One of the things I worry about is people getting captured by their tools," said Jane Street CTO Ron Minsky on the firm's podcast, Signals and Threads, this week. Jane Street has used functional language OCaml for decades and made sure that everyone knew the language so that they "can go anywhere into any code base and do stuff." Thanks to Python, Minsky said there has been a "natural striation where people get good at the one thing," but aren't necessarily plug-and-play. Minsky said "I'd really love to be more in a world where there are no OCaml programmers and Python programmers... just software engineers."
Unlike OCaml, Python is not a niche language. In fact, it's the most popular language there is, topping the TIOBE index. "Everyone knows Python," Minsky said, "but there are lots of people who aren't any good at it and there's real things to learn to be good at it." He said there are also "people who've been [at Jane Street] for a long time," who may not have had the need to learn Python until very recently.
Aaron Bauer, a developer educator at Jane Street, said on the podcast that a lot of the work researchers and traders do in Python is via notebooks: "there are different notebook solutions [but] they all have different significant flaws or limitations." One common limitation is that the code in these notebooks is "not always easily shareable," but Bauer said Jane Street is building a solution to fix this.
It's not just the languages changing at Jane Street, it's the nature of the coding process. Minsky said that "the modern condition of a software engineer" involves "a lot more [code] review as a proportion of your time" thanks to agentic coding tools. Minsky said it's a "real problem" when engineers "vibe it up too much and use LLMs to generate stuff and do not think." He said that AI generated programs can be "uncanny" in that it can look well written and presentable, have very few inidications that there would be a problem with the code and yet "they're often still super broken."
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.
