Jane Street Details How AI Raises the Stakes for Trading Data Work

Jane Street has laid out how LLMs and stronger machine-learning models are changing trading research: they make messy external data more useful, but also make provenance, timestamps and human judgment more important. In a September 2 Signals and Threads episode, Eric Mannes, who helps lead the firm's alternative data team, described a workflow built around preserving what data was available at the time, not just what later corrections made accurate.