Applied AI Scientist, Simulation & Reinforcement Learning

Role overview Own and improve a discrete-event simulator used to develop and evaluate systems that coordinate people and robots in warehouse operations. The role begins with strengthening simulator fidelity, reliability, and experiment tooling, then grows to include training and assessing reinforcement-learning policies against existing approaches and real operational data.

Responsibilities - Develop and maintain a SimPy-based simulator, improving its code quality, tests, reproducibility, and supporting tools. - Build infrastructure for running, managing, and reporting experiments across warehouse scenarios and benchmark sites. - Compare simulation results with operational data, tune parameters, investigate anomalies, and reduce differences between simulated and real behavior. - Represent operational dynamics such as congestion, worker behavior, and queues at an appropriate level of detail. - Create a Gym-style environment, train and evaluate learning-based policies, and compare their performance with current heuristics. - Communicate results and uncertainty to engineering, product, and operations partners; assess which research methods are suitable for deployment.

Requirements - A PhD in a relevant quantitative field, or an MSc with at least five years of industry experience applying AI, machine learning, or simulation to real problems. - Hands-on experience building, extending, or validating simulators; discrete-event simulation experience is especially relevant. - Practical reinforcement-learning experience, including training agents in simulation and working with Gym-style environments. - Strong Python and software engineering skills, including testing, code review, maintainability, and reproducibility. - Sound experimental and statistical practice, including benchmarking, tuning, ablation studies, and handling uncertainty. - Clear English communication and experience in robotics, logistics, warehousing, or another operations-heavy field.

Nice to have - Research publications in AI, machine learning, robotics, or computer vision venues. - Experience with optimization methods or tools such as Optuna, multi-agent systems, or production ML and MLOps.

Benefits and work setup - Remote role for candidates residing in the United Kingdom who already have the legal right to work there; visa sponsorship is unavailable. - Published base salary range of GBP 70,000–120,000, with bonus and equity.