Research Scientist, Simulation

The role

You will build simulated worlds that power loco-manipulation policies for a wheeled mobile manipulator. You will own the end-to-end pipeline from real-to-sim reconstruction to procedural scene generation and sim-based training data.

What you will do

• Build a pipeline that reconstructs real, facility-scale environments into simulation.
• Procedurally generate navigable environments to stress-test loco-manipulation policies.
• Generate synthetic training data for policies that jointly reason about base positioning and arm manipulation.
• Build simulation benchmarks that test joint base and arm policies.
• Push photorealistic rendering quality to close the sim-to-real gap.
• Partner with the AI Research team on simulated data and evaluation.
• Partner with Data Ops on real-world vs. simulated data capture.

Required qualifications

• MS or PhD in CS/Robotics/Graphics, or equivalent hands-on experience.
• Hands-on experience with simulation stacks (MuJoCo, Isaac Sim/Isaac Lab, SAPIEN, Omniverse, Blender, or similar) for robotics or graphics.
• Experience with procedural scene/asset generation, domain randomization, or photorealistic rendering pipelines at scale.
• Familiarity with training/evaluating manipulation or loco-manipulation policies (imitation learning, VLA, diffusion, or RL).
• Strong Python skills.
• Comfort with PyTorch or JAX for model training/eval.
• Senior enough to define your own research agenda and exercise independent judgment.

Nice to have

• Experience simulating mobile bases (wheeled, tracked, or holonomic).
• Prior research or engineering background in loco-manipulation or whole-body control.
• Experience with real-to-sim-to-real pipelines (3D reconstruction, NeRF/Gaussian splatting, differentiable rendering).
• Exposure to world-model / video-prediction research.
• GPU-scale physics simulation experience (CUDA, large-batch parallel sim).
• Background simulating wheeled/mobile manipulators specifically.
• Publications at CoRL, RSS, ICRA, NeurIPS, CVPR, or SIGGRAPH.
• Experience leading or mentoring within a small research team.

About the company

Dyna Robotics builds general-purpose robots powered by a proprietary embodied AI foundation model. The team comes from Google DeepMind, Meta, and Cruise.