Anomaly Detection in High-Performance Trading Systems
January 1, 2021Tradeweb Markets Inc. Computer Science, 2020–21
Liaison(s): Justin Peterson ’85, Stefan Kutko
Advisor(s): Elizabeth Sweedyk
Students(s): Julia Puzzo (PM), Brandon Theodorou, Riti Sharma, Jacob Donenfeld
Tradeweb is the leading trading platform for fixed income products, derivatives, and ETFs. The operations of Tradeweb’s systems are dependent on trading latency being below a certain threshold, as high latency can negatively impact client experience. The Tradeweb Clinic Team is using machine learning models to produce a tool that predicts latency spikes by monitoring the underlying operational system, so that Tradeweb can be made aware of and take quick action to resolve any sub-optimal performance.