Applied Scientist, Robotics Simulation

Collinear AI, Inc.; San Francisco (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

Collinear AI, Inc. seeks an Applied Scientist (Robotics Simulation) to design the physics‑solving engine behind its synthetic data platform. You will architect a next‑gen simulation stack for maximum throughput, scalable environments, and high physical fidelity.

You will collaborate with GPU engineers, scene generation teams, and ML researchers to push model training with synthetic data. Expertise in physics engines and parallel solvers is essential.

Qualifications

• Advanced degree in a relevant field or equivalent hands‑on industry experience in physics‑based simulation.
• Strong track record developing/high‑fidelity physics engines and solvers.
• Deep understanding of mechanics, contact modeling, constraints, integrators, and large‑scale trade‑offs.
• Expertise in CUDA and GPU optimization for parallel numerical algorithms.
• Proficiency in C++ and Python for high‑throughput production software.
• Experience with differentiable simulation and ML data pipelines.

Responsibilities

• Physics Solver Innovation: develop advanced solvers for high‑DoF, contacts, deformables, and fluids.
• Training Data Optimization: design simulation environments and domain randomization for high‑utility datasets.
• Procedural Scene & Scale Integration: scale environment variations and multi‑agent interactions in billions of steps.
• Performance Profiling: profile and optimize multi‑GPU rollout pipelines with low latency and memory overhead.
• Technical Leadership: shape long‑term roadmap for high‑fidelity physical modeling and synthetic data engines.

Skills

Physics simulation

CUDA

C++

Python

High-performance computing

Differentiable simulation

Education

MS in Physics

MS in Computer Science

PhD in Physics

PhD in Computer Science

Tools

CUDA

C++

Python

Job description

Collinear AI, Inc. seeks an Applied Scientist (Robotics Simulation) to design the physics‑solving engine behind its synthetic data platform. You will architect a next‑gen simulation stack for maximum throughput, scalable environments, and high physical fidelity.

You will collaborate with GPU engineers, scene generation teams, and ML researchers to push model training with synthetic data. Expertise in physics engines and parallel solvers is essential.