Peter Efstathiou

Background

About Peter Efstathiou

Expert with Python for machine learning, data analysis, and grid computing for large datasets. Fascinated with transaction cost modelling (TCA), understanding market impact, and developing optimal execution models. Strongly influenced by transient impact propagator models. Delighted by understanding and predicting details of market microstructure. Bayesian in approach. Particularly influenced by Jayne’s Logic of Science and Inductive Logic Programming, which explore the limits of what it is possible to learn, either with small number of samples or extremely low signal to noise, and then how to make decisions given the meta-uncertainty. Detailed knowledge and practical experience of machine learning (e.g. Elements by Hastie et al). Sophisticated user of Bloomberg terminal and API. Extremely proficient in Excel, SQL (MSSQL & postgresql), git, C#, Mathematica, and the Linux OS. All round modern software experience with a wide range of languages, SQL and timeseries databases, project management software, and modern software devops and best practices. Regularly read academic and technical literature extensively and critically, often reproducing results.