Events

MIC seminar: Training Foundation Models for Algorithmic Trading

Speaker: Prof. Zhangyang "Atlas" Wang from UT Austin

Location: 60 Fifth Avenue, Room 7th floor open space

Date: Friday, April 18, 2025

In this talk, I will introduce the pioneering efforts of XTX Markets in applying large-scale foundation models to algorithmic trading. I will highlight both the unique challenges and exciting opportunities in forecasting financial markets using sophisticated deep neural networks trained on massive, heterogeneous datasets. Specifically, I will present a high-level discussion of how our models can be trained to utilize diverse market data to generate robust forecasts across thousands of financial instruments globally. I will also cover our experiences in "neural scaling laws" on market data as a novel modality, amid extreme market noise and volatility. Finally, I will introduce our latest research initiatives, emphasizing XTX's substantial investments in computational infrastructure, innovative modeling approaches, and proprietary research.