Research Engineer / Research Scientist — Robotics & Physical AI

ABOUT THE ROLE

Join a small, high-impact team at an early-stage, Y Combinator–backed AI research company building the data and learning systems that enable robots to perceive, reason, and act in the physical world. This is a rare opportunity to work across the full stack — from learning algorithms and simulation to real hardware — and to do research that ships into real-world deployments.

The team operates at the intersection of Physical AI, embodied intelligence, and high-quality training data, collaborating with leading robotics companies and research labs. All levels of experience are welcome, from new graduates to seasoned researchers and engineers.

WHAT YOU'LL DO

• Develop learning systems for robots, including policies, world models, and vision-language-action (VLA) models.
• Build and maintain simulation environments and evaluation frameworks for robotic tasks.
• Apply and advance computer vision and perception algorithms for autonomous robots.
• Build pipelines connecting simulation and real-world robot deployment (real-to-sim and sim-to-real).
• Work hands-on with robot hardware, sensors, actuators, and embedded systems to prototype and validate capabilities.
• Train and evaluate reinforcement-learning (RL) and imitation-learning systems for manipulation, locomotion, or mobile platforms.
• Integrate and experiment with robot platforms to validate algorithms in the physical world.
• Create internal tools and infrastructure that accelerate robotics R&D.
• Collaborate with external robotics companies and research labs on technical challenges.

WHAT WE'RE LOOKING FOR

• Eligibility to work in the United States without employer sponsorship (visa sponsorship is not available).
• Ability to work on-site in San Francisco — this role is not remote.
• Strong proficiency in Python and C++.
• Hands-on experience with robotics hardware, sensors, actuators, and embedded systems.
• Practical experience with deep learning, reinforcement learning, and computer vision.
• Experience with real-to-sim and sim-to-real workflows.
• Experience with embodied AI, RL for robotics, or vision-language-action models.

Nice to Have

• Familiarity with simulation platforms such as Isaac Lab, MuJoCo, MJX, Genesis, Gazebo, or PyBullet.
• Experience with CAD tools or Blender for asset creation.
• Exposure to modern ML methods: LLMs, Transformers, multimodal AI, diffusion models, or AI agents.
• Ability to communicate and collaborate effectively with robotics companies and external research labs. Mindset: You're comfortable moving fluidly between research, engineering, experimentation, and product development in a small, fast-moving team. You want to ship, not just publish.

COMPENSATION

& BENEFITS

• Salary: $100,000 – $300,000 USD annually, depending on experience and seniority.
• Early-stage equity opportunity at a seed-stage company backed by a top accelerator. LOCATION
• On-site in San Francisco, CA, USA. Remote work is not available for this role.
• Visa sponsorship is not available; candidates must be authorized to work in the US.