
Sign up to save your podcasts
Or
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.
You can find the transcript for this episode on our website.
Some links to topics that came up in the discussion:
4.9
7272 ratings
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.
You can find the transcript for this episode on our website.
Some links to topics that came up in the discussion:
272 Listeners
582 Listeners
584 Listeners
2,384 Listeners
624 Listeners
1,842 Listeners
191 Listeners
227 Listeners
82 Listeners
424 Listeners
58 Listeners
71 Listeners
379 Listeners
122 Listeners
63 Listeners