Overview

Usman’s research interests lie in compositional and hierarchical methods for improving the generality, explainability and flexibility of learning systems, particularly within Deep Reinforcement Learning (RL), where e.g. tasks are made up hierarchically of composable and trasferrable subtasks. Before commencing his PhD program, Usman worked in industry as a research engineer, developing new methods for safe RL and autonomous driving. He also has several years’ experience as a software engineer, building backend services and automation tools for Amazon Prime Video. Prior to starting his career, Usman completed his Master’s in Mathematics at Imperial College London and has a particular preference for mathematically involved works as well as those that take inspiration from human intelligence.

Education

• Master of Science (Integrated) MSci - Mathematics, Imperial College London