The Uncertain Art of Accelerating ML Models with Sylvain Gugger
About this episode
Sylvain Gugger discusses the intricacies of accelerating machine learning models in trading environments.
Sylvain Gugger is a former math teacher who fell into machine learning via a MOOC and became an expert in the low-level performance details of neural networks. He’s now on the ML infrastructure team at Jane Street, where he helps traders speed up their models. In this episode, Sylvain and Ron go deep on learning rate schedules; the subtle performance bugs PyTorch lets you write; how to keep a hungry GPU well-fed; and lots more, including the foremost importance of reproducibility in training runs. They also discuss some of the unique challenges of doing ML in the world of trading, like the unusual size and shape of market data and the need to do inference at shockingly low latencies.
More episodes of Signals and Threads
- Why Testing is Hard and How to Fix it with Will Wilson
- Why ML Needs a New Programming Language with Chris Lattner
- The Thermodynamics of Trading with Daniel Pontecorvo
- Building Tools for Traders with Ian Henry
- Finding Signal in the Noise with In Young Cho
- Solving Puzzles in Production with Liora Friedberg
Explore listener stats, chart rankings, contacts and more on the Signals and Threads podcast page.