Applied AI Scientist – Simulation, Reinforcement Learning

This is a fully remote position, open to applicants in United Kingdom.

📋 Description

• Take ownership of and enhance the SimPy-based discrete-event simulator.

• Elevate test coverage, code quality, reproducibility, and developer tools.

• Develop infrastructure for experiments, including orchestration, study management, reporting, and automated benchmark-site operations.

• Create and sustain site data required to replicate customer warehouse environments.

• Facilitate scenario testing and sensitivity analysis for the order-engine and orchestration teams.

• Identify, quantify, and minimize the sim-to-real gap by comparing with actual site data.

• Conduct parameter optimization and tuning studies while diagnosing anomalies and non-deterministic behavior.

• Model warehouse operations, including congestion patterns, picker behavior, and queue dynamics.

• Convey results and insights confidently to engineering, product, and operations stakeholders.

• Transform the simulator into a reinforcement-learning environment akin to Gymnasium/Gym.

• Design, train, and assess reinforcement learning policies against heuristic algorithms.

• Mitigate overfitting to simulator artifacts and ensure effective transfer to real-world applications.

• Prototype, assess, and implement learning-based components in collaboration with ML, MLOps, and OE teams.

• Stay updated with advancements in reinforcement learning, multi-agent systems, and simulation research.

• Contribute to technical papers, patents, and relevant publications.

• Report directly to Oscar Mendez within the Applied Data Science team.

⛳️ Requirements

• PhD in Computer Science, AI/ML, Robotics, Mathematics, Physics, Operations Research, or a related quantitative discipline, or a Master's degree with 5+ years of industry experience applying AI/ML or simulation to real-world challenges.

• Hands-on experience in creating, extending, or validating simulators, preferably in discrete-event simulation (SimPy or equivalent).

• Practical experience with reinforcement learning and training simulation-based agents, including simulation for RL and Gym-style environments.

• Strong foundation in Python and software engineering principles: testing, code review, and developing maintainable and reproducible code.

• Rigorous experimental and statistical mindset: benchmarking, tuning, ablation studies, and handling uncertainty.

• Background in robotics, logistics, warehouse, or other operations-intensive fields.

• Publications in notable AI, ML, robotics, or vision conferences.

• Exceptional English communication skills, both written and verbal.

• Legal authorization to work in the UK.

• Must be a resident of the UK.

• Visa sponsorship is not available.

🏝️ Benefits

• Bonus opportunities.

• Equity participation.

• Continuous learning and career advancement prospects.

• Chance to develop technology utilized on a global scale.