Manipulation Capabilities Engineer

📋 Description

• Post-train manipulation policies via behaviour cloning and RL from data to deployment.
• Data preprocessing strategies to improve data quality.
• Set up RL training with a digital twin; iterate reward and sim quality for real-world transfer.
• Drive data collection to define good data, ensure diversity and coverage.
• Expand observation and action spaces for new capabilities; work with Teleoperations to expose to operators.
• Partner with Teleoperations and Controls to improve motion smoothness.

🎯 Requirements

• 3+ years on robots with shipped artifacts.
• Experience with neural network post-training.
• Deep learning infra: streaming data, checkpointing, distributed training, PyTorch or JAX.
• Strong software engineering practices.
• Clear experiment documentation and trade-off communication.

🎁 Benefits

• Competitive equity: stock options with upside.
• 30+ paid days off, including 23 days annual leave and UK holidays.
• Private healthcare, virtual and in-person care.
• Pension scheme with 8% total contribution.
• Free daily breakfast, catered lunch, and snacks in-office.
• Work at the frontier with world-class engineers and researchers.