Senior ML Engineer – AI Research, Physical AI

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

📋 Description

• Conduct applied research in Physical AI aimed at creating intelligent agents that can perceive, reason, and act within the physical environment.

• Modify extensive foundation models and learning algorithms tailored for robotic agents.

• Prototype innovative capabilities in simulations and validate these approaches on actual systems.

• Design, implement, train, and assess large models and learning algorithms.

• Develop architectures that integrate vision, language, and action, linking multimodal perception and language comprehension with physical control.

• Explore reinforcement and imitation learning techniques for challenging objectives.

• Create scalable methods that incorporate demonstrations, teleoperation data, video, simulation trajectories, and autonomous robot experiences into foundation models.

• Design methodologies for data capture, datasets, evaluation protocols, and data-quality pipelines tailored for embodied learning.

• Develop simulation environments and execute sim-to-real experiments on physical robotic platforms.

• Investigate planning, guided generation, and search strategies over action trajectories.

• Prototype capabilities for dexterous manipulation, mobile manipulation, and whole-body control.

• Develop robust research software and distributed training infrastructures.

• Collaborate with research, infrastructure, and engineering teams to transform ideas into dependable real-world systems.

• Communicate findings through technical reports, open-source releases, demonstrations, and research publications.

⛳️ Requirements

• A deep understanding of the theoretical principles underlying machine learning, reinforcement learning, or robot learning.

• Extensive expertise in at least one relevant domain, such as reinforcement learning, imitation learning, multimodal generative modeling, computer vision, robotics, planning, or control.

• Experience in training and evaluating contemporary deep learning models, including transformer-based or multimodal foundation models.

• Significant experience in training large models across various computational nodes.

• Strong software engineering and algorithm design capabilities, primarily utilizing Python.

• In-depth experience with JAX.

• Proven track record in designing, executing, and analyzing machine learning experiments with appropriate statistical rigor.

• Ability to formulate meaningful research questions, design experiments to test clear hypotheses, and derive defensible conclusions.

• Experience in implementing research concepts and iterating swiftly across modeling, data, infrastructure, and evaluation.

• Strong communication and leadership skills, with the ability to collaborate across research and engineering disciplines.

• Capability to document research findings clearly and contribute to technical reports or academic publications.

• Applicants must have authorization to work in the country they are applying to and provide proof of employment eligibility as a condition of hire.

• Nice-to-have: experience with real-world robots and robotic simulation environments.

• Nice-to-have: experience with dexterous, whole-arm, mobile, or humanoid robotics.

• Nice-to-have: experience with multimodal sensing.

• Nice-to-have: experience in collecting human demonstrations.

• Nice-to-have: experience in developing or post-training vision-language, vision-language-action, video, or world models.

• Nice-to-have: experience with deep reinforcement learning techniques such as offline RL, actor-critic methods, PPO, reward modeling, preference learning, or model-based RL.

• Nice-to-have: familiarity with MuJoCo, Isaac Sim, Isaac Lab, PyBullet, ROS, or equivalent systems.

• Nice-to-have: knowledge of FSDP, ZeRO, FlashAttention, mixed-precision training, quantization, and distributed checkpointing.

• Nice-to-have: PhD in Computer Science, Robotics, Machine Learning, Artificial Intelligence, or a related technical field, or equivalent practical experience.

• Nice-to-have: a record of impactful publications, open-source contributions, or deployed robotic systems.

• Nice-to-have: experience in engineering large distributed data-processing, simulation, or model-training systems.

• Nice-to-have: a history of building and delivering products or research prototypes.

• Nice-to-have: excellent command of English along with strong technical writing, presentation, and communication skills.

• Nice-to-have: proficiency in version control, testing, code review, and CI/CD.

🏝️ Benefits

• Competitive compensation.

• Opportunities for career growth and learning.

• Flexibility and ownership in your work.

• A collaborative and innovative culture.

• The chance to work on impactful AI projects.

• An international environment with talented teams.

• A fast-paced work setting.

• Encouragement of bold thinking.

• Continuous growth opportunities.

• The potential for meaningful impact.

• Trust and genuine ownership in your role.

• The opportunity to shape the future of AI.