AI is changing careers in quant finance, and not just for your average quant in a bank. Electronic trading firms and market makers have spent years hoovering up elite talent, and AI is making them even more productive. It might also make that same talent more likely to leave…
Click here to join the bubble by eFinancialCareers, our new anonymous community. ✍️
How do electronic trading firms use AI?
AI is a broad field and quant firms have been using it for a long time. Machine learning tools like natural language processing (NLP) and computer vision have been used by quants for decades; they allow quants to look for specific signals in large datasets that tell them whether to buy or sell an asset. NLP, for example, can analyze millions of social media posts and search for specific keywords, grammar or other signals to extract broad consumer sentiment on a specific stock. Reinforcement learning, a technique where an AI model uses simulated trial and error to assess the optimal course of action, can be used to develop execution algorithms that make trades at the most efficient time. Similar techniques can be used to analyze post trade data in order to improve the efficiency of models going forward. They can also be used by quant risk teams to ascertain how much capital should be given to each trader.
In 2026, however, AI is reaching a whole new level in electronic trading firms. Jane Street has encouraged traders to build trading strategies using Python since at least 2024. It says it’s particularly fond of the PyTorch library (instead of other options like TensorFlow) as it allows traders to iterate through different strategies very quickly. Burgeoning languages like Mojo are also under consideration.
As AI takes over, the nature of the work has not changed, but the volume of data and speed of analysis has gone parabolic. The GPUs facilitating the analysis have become more powerful, and large language models can ingest and structure data much more efficiently, meaning that that traditional machine learning methods can be applied to larger datasets for richer analysis.
Many trading firms have massively grown their personal GPU clusters which facilitate their machine learning strategies. Jane Street seems to have doubled its number of GPUs in the past year and now has tens of thousands of them. Quadrature, a smaller AI trading firm paying millions per head, had ~20,000 GPUs in November, equivalent to roughly 11 chips per employee. XTX Markets has one of the most impressive clusters; open job listings say the market maker has “25,000 GPUs with 650 petabytes of usable storage.”
Trading firms have rapidly adopted AI tools while also building their own. At Hudson River Trading, staff are spending $1k a day on tokens and are being celebrated for it. Jane Street, meanwhile, has had difficulty adopting off-the-shelf coding tools because it uses OCaml, a relatively niche language; John Crepezzi, a Jane Street engineer who develops its agentic tools, said in a talk last year that there’s more OCaml code in Jane Street than there is in the rest of the world combined.
Trading firms have made massive investments in AI beyond their traditional business models by building out venture capital arms. Jane Street, for example, has a stake in Anthropic. Investments by quant firms aren’t AI specific, though; Jump Trading and Susquehanna have both invested in prediction markets platforms like Kalshi and Polymarket.
Which roles in trading firms are safest from AI? Which are most under threat?
Annanay Kapila, a former high-frequency trader who now runs startup exchange QFEX, told us that “the best people will harness AI very well, become super productive and make tonnes of money.” Roles in research, infrastructure and development are already starting to converge as firms want talent who can fluidly work across different domains rather than becoming siloed.
Annanay said that “genuine, idiosyncratic human alpha will become more elevated and more productive.” Traders who incorporate human-driven decision making are “probably the most protected” while purely systematic modelling roles are under threat. “[Claude] Fable has become really good at quant research,” Kapila said. “It’s able to spin up all the basic pieces of a trading strategy very quickly and identify a bunch of patterns.”
People who source data for their firms will also be safe, if not more valuable. “Getting access to data faster or getting access to data that other people don’t have is going to become really important,” Kapila said. Everyone has access to the Bloomberg terminal, but these professionals are focused on obtaining obscure data like “video footage outside French power plants checking how much smoke is going out of the chimney.”
In most industries, junior roles are being hurt the most by AI. This is not true for junior quants in market making firms, where the bar has always been high. “Unlike engineering, you can’t make an impact as a junior unless you’re really sharp,” Kapila said. Junior quants use AI to speed up the process of modelling and writing trading algorithms, allowing them to test more ideas more quickly.
As with any job that involves coding, experience with AI coding tools is fundamental. However, a senior quant recruiter, speaking anonymously, told us that “one of the biggest concerns for elite trading firms right now is candidates becoming too dependent on AI and losing their core coding and problem solving abilities.”
The talent war between trading firms and AI labs
The biggest impact on jobs in electronic trading firms is less about AI tools and more about the firms building them. AI labs like Anthropic and OpenAI have been open about their pursuit of talent from high-frequency trading firms. In 2024, $3m compensation packages were offered to HFT quants with just a few years of experience. Insiders say that there has been a significant brain drain from quant finance to AI, and that there hasn’t been much traffic in the opposite direction.
Today, that same pool of talent is still valuable. Anthropic openings for research engineers which request experience in quant finance (or at a rival AI lab) offer up to $850k in salary alone. You’ll often get paid much more in equity, which can balloon in value as the labs continue to grow. One OpenAI employee told Business Insider last week that they own $50m in OpenAI stock despite working there for just three years due to its swelling valuation.
For many candidates, the money isn’t even the issue. Candidates we’ve spoken to have said they prefer working in AI labs because of the opportunity to feature in academic publications, and an alignment with the mission of reaching AGI. In quant funds, much of your best work will be kept proprietary.
How does the work compare? Grant Stenger, an ex-Jane Street trader speaking to FT Alphaville earlier this year, said that “the core job is actually identical;” you’ll use data to build a model, execute it within a constrained environment, and use feedback to enhance the model in future iterations. In terms of culture, quants say you tend to work fewer hours in finance as your work revolves around the markets, but AI lab staff say that the long hours are a good thing for them because of their commitment to the cause.
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.
Photo by Arturo Añez on Unsplash