30초 요약
Matt, a trader on Jane Street’s Hong Kong equities desk, discusses applying machine learning to financial data and the infrastructure needed to process it. He also describes collaboration among traders, researchers and developers, and the distinction between ML engineering and research roles.
- 1A smaller office, broader exposureMatt says Hong Kong’s smaller office gives people exposure to different areas and often more room to take responsibility.
- 2Financial data changes the problemFinancial data is often noisier than other kinds of data and exists in large quantities. Jane Street has built infrastructure to process it, using Python and OCaml.
- 3Collaboration across rolesMatt sits with researchers and developers, and colleagues often gather around a desk to discuss research.
- 4Two machine learning tracksML engineers build infrastructure and tools; ML researchers use those tools to conduct research. Matt says the firm is interested in candidates with prior ML experience.
핵심 정리
- The smaller Hong Kong office can offer exposure to different areas and opportunities to take responsibility.
- Financial data is often noisy and plentiful, requiring techniques and infrastructure suited to its properties.
- Jane Street’s research infrastructure uses Python extensively alongside OCaml.
- Traders, researchers and developers collaborate closely on research.
- ML engineering focuses on infrastructure and tools, while ML research focuses on using those tools for research.
챕터별로 다듬은 전문 읽기
From New York to Hong Kong
Matt introduces his trading role and move to Hong Kong. He values the exposure and responsibility available in a smaller office.
I’m Matt, a trader on the Hong Kong equities desk. I started in New York in 2012 and moved to Hong Kong a couple of years later. Since Jane Street started in New York, that office is much bigger; in Hong Kong, I get more exposure to different areas and often more room to take on responsibility, which I find exciting.
Machine learning on financial data
Matt explains how noisy, abundant financial data shapes the work and the tools it requires. He describes the firm’s growing ML effort.
The main difference between working on machine learning at Jane Street and at some large technology companies is the class of problems. We focus on financial data rather than, for example, language data. Financial data has different properties: it is often noisier, which makes the techniques somewhat different.
There is also a huge amount of financial data, so we need a lot of tooling to process it. We have built strong infrastructure for that. Most of our research infrastructure is in Python, though we also do a lot in OCaml.
We’re excited to grow our machine learning efforts and expand the team globally. It’s an exciting time to join because this is one of our fastest-growing areas.
Collaboration and ML careers
Traders, researchers and developers work together. Matt outlines the engineering and research tracks and explains what he enjoys about the work.
Jane Street has a collaborative culture. People from different groups sit together: I’m a trader, but I sit with researchers and developers, and we collaborate. It’s common to see several people gathered around someone’s desk discussing interesting research.
We welcome candidates from a variety of backgrounds, but for ML roles in particular we’re interested in people with some prior ML experience. We have two tracks: ML engineering focuses on infrastructure and building tools, while ML research focuses on using those tools to do research.
My favourite part of working here is probably my co-workers. I also find the problems we work on very interesting. Sometimes I work on them at the weekend, not because anyone asks me to, but because I’m interested in them.
같은 채널의 다른 글
15:33EN
Jane Street· 2026년 10월 26일Figgie Strategy: Reading Hands, Winning Majorities and Pricing CardsBradley, Mike · Jane StreetJane Street traders Bradley and Mike analyze three games to show how private information, market competition and timing affect card prices.
1:48:29ENJane Street· 2026년 10월 16일How AI Is Changing Programming: John Crepezzi and Ron MinskyYaron Minsky, John Crepezzi · Jane StreetJane Street’s AI tooling lead and the Signals and Threads host examine agent infrastructure, security, code review and the changing experience of writing software.
1:36:37ENJane Street· 2026년 9월 23일Aaron Bauer on Teaching Software Engineering in the Age of AIYaron Minsky, Aaron Bauer · Jane StreetHow learning goals, active practice and editor design shape engineering education when models can generate the code.
1:05:45ENJane Street· 2026년 9월 11일Cyrus Omar on Live Programming and a Computational Commons for Humans and AICyrus OmarA demonstration of Hazel shows how immediate feedback, structured collaboration and carefully scoped permissions could support people and AI agents working together.
1:25:00ENJane Street· 2026년 9월 2일Eric Mannes on Turning Messy External Data Into Trading ResearchYaron Minsky, Eric Mannes · Jane StreetFrom commodity futures and an Excel risk system to point-in-time datasets and AI, Eric Mannes explains the domain knowledge behind data engineering at Jane Street.
1:16:29ENJane Street· 2026년 8월 26일Grant Sanderson and Alok Puranik: Positional Encodings Through Group TheoryGrant Sanderson, Alok Puranik · 3Blue1BrownStarting with assumptions about attention, they derive matrix exponentials and connect rotations, decay and Jordan blocks to RoPE and ALiBi.