Research Engineer / Scientist (Robot Learning)

World Labs•San Francisco, CA

13d

About The Position

World Labs is building foundational world models that can perceive, generate, reason, and interact with the virtual and physical worlds, unlocking AI's full potential through spatial intelligence. This role is for a Robot Learning Engineer/Scientist focused on developing and advancing state-of-the-art methods for training end-to-end robot policies, with an emphasis on sim-to-real transfer, robust performance, and scalable training and inference pipelines. It's a hands-on, research-driven position at the intersection of robotics and machine learning, requiring close collaboration with research scientists, ML engineers, and system teams to translate robotic policies into production-ready systems.

Requirements

6+ years of experience working on manipulation, locomotion, robot policy training, or related areas.
Strong foundation in robotics, neural network designs, and sim-real transfer.
Deep experience with robot policy designs (e.g., VLA, WAM, diffusion).
Proficiency in Python and/or C++, with hands-on experience building research or production robotic systems.
Experience with deep learning frameworks (e.g., PyTorch) and low-level robotic controllers.
Proven ability to work in ambiguous, fast-moving environments and drive projects from concept through deployment.
A strong sense of ownership and engineering rigor: care deeply about correctness, stability, and measurable improvements.
Enjoy collaborating with a small, high-caliber team and raising the technical bar through thoughtful design, experimentation, and code quality.

Responsibilities

Design and implement modern robot learning systems, including imitation learning and reinforcement learning for manipulation.
Research, prototype, and productionize robotic policies with a focus on speed, precision, and scalability.
Develop and improve training pipelines for sim-to-real transfer, including domain randomization, system identification, and real-sim alignment.
Collaborate with simulation and infrastructure teams to minimize the sim-to-real gap and ensure learning methods integrate cleanly with real-robot deployment stacks.
Build end-to-end training and evaluation workflows for robot policies, from large-scale data generation to scaling up training and evaluation.
Optimize policy performance across the stack, including training speed, inference latency, and data generation efficiency to support large-scale production constraints.
Diagnose failure modes in simulation and real-world rollouts, and design principled solutions to improve robustness, efficiency, and generalization.
Contribute to technical direction by proposing new research ideas, mentoring teammates, and helping set best practices for robot learning across the organization.