RAI AI Robot Learning Engineer

About REVELREVEL is an AI robotics company developing physical intelligence for general-purpose humanoid robots. We capture the force, dexterity and intent of human work with our Neural Gambit wearable, and use it to train RAI, the intelligence that powers our robots. REVEL is headquartered in Palo Alto, California, with R&D and engineering facilities in Prague and Hradec Králové, Czech Republic. This role is on-site with our engineering team in the Czech Republic.Develop imitation- and reinforcement-learning methods that turn demonstrations into skills. Our RAI team builds the core intelligence that powers REVEL robots, learning from the force and intent Neural Gambit captures.ResponsibilitiesTrain and evaluate models on REVEL's human-demonstration dataBuild scalable, reproducible training and evaluation pipelinesCollaborate across the RAI, hardware and data teamsPush results from research into production on the robotRequirements3+ years of relevant professional experienceHaving modified robot-learning models and taken them from training through on-robot deployment, inference, testing, and validation—not only used off-the-shelf implementationsExperience building and shipping robot-learning systems on real robots, not only in simulationExperience with one or more of the following: behavior cloning, imitation learning, VLAs, VLMs, Transformers, Diffusion Policies, or world modelsProficient in Python and/or C++ and deep-learning frameworks such as PyTorch for building training pipelines, fine-tuning models, and developing robot-facing codeComfortable designing experiments, diagnosing failures, and iterating quickly on physical systemsBonus qualificationsExperience using reinforcement learning approaches—such as PPO, SAC, offline RL, or RL fine-tuning—to improve learned robot behaviorsTrack record of deploying learning-based manipulation systems on commercial or production robots, not only demosPrior work with humanoids or other highly dexterous robotic platformsPublication record in robot learning, manipulation, or embodied AIBonusesWork That Ships: We capture how skilled humans work, their force, touch and judgment, and our robots do the work. You put robots on a paying customer's floor, not in a demo loopThe Team: Colleagues from NVIDIA, SpaceX and Neura Robotics, and founders you work with directly. No layers, no process between you and the decisionsEquity for Key Roles: For select positions, meaningful stock options mean you're not just working here, you own a piece of the outcomeSalary and Quarterly Bonus: Strong base pay plus a quarterly bonus, in a city where it goes furtherUnlimited Paid Time Off: Real flexibility to take time away when you need it. We trust our people to own their work, their time and their resultsYour Own Hardware: A top-spec GPU workstation, cluster access, and hands-on time with the robots you're building. Not a software sandboxKeep Learning: Conference budget for select roles (GTC, CoRL, ICRA), and room to publish and contribute to open source where our IP allowsPrague, On-Site: Robots need hands, so we work together in our Prague lab. Moving here? We sponsor your work visa and cover relocationHealth and Fitness: Multisport card (from December 2026), extra paid sick days, and an employer pension contributionLunch, On Us: Complimentary lunch every working day, plus coffee, snacks and drinks whenever you need a boostHow to applyApply through the link on this posting. Include a short description of an environment you administered or took over: its size, what you changed about how access or devices were managed, and what you would do differently now.