The competitive environment makes machine learning in trading fundamentally different from other applications of ML. Quantitative Researcher Denis Perevalov shares what makes it interesting.
NeurIPS Conference
ConferenceGlobal
Financial markets present a distinct ML challenge: noisy, non-stationary data in a competitive and adaptive environment. Models that perform well on historical data can behave differently in live trading. Meet our researchers at NeurIPS to explore the research problems we're working on and how we approach them.
From hypothesis to live markets
ML is applied to price prediction and execution problems across thousands of instruments and a range of time horizons. Teams use deep learning on raw order-flow data, LLMs to extract predictive signals from financial news and reinforcement learning to improve decisions about when and how we trade.
Research in numbers
- 0B+
predictions daily
- Petabytes
of market data
- μs
between market orders
