Genesis-World: Core Simulation Engine Engineer

Genesis•San Carlos, CA

12h

About The Position

Genesis-World is an open-source, general-purpose simulation platform for physical AI from Genesis AI. It features a unified multi-physics engine supporting rigid bodies, FEM, MPM, particles, cloth, and fluids, along with an in-house renderer, Nyx, for realistic rendering. The platform supports a wide range of sensors, including cameras, lidar, IMU, contact forces, and temperature, as well as advanced tactile simulation. Genesis-World is developing a comprehensive and fast Incremental Potential Contact solver for deformable body dynamics. It powers real business applications and is Python-first, running on various hardware architectures through an in-house JIT compiler called Quadrants. The platform aims to accelerate robotics development by providing a simulation environment that runs significantly faster than real-world hardware while maintaining high correlation with on-hardware results. The role focuses on improving the reliability, maintainability, and user-friendliness of the Genesis-World codebase, addressing friction points in simulation usage and debugging.

Requirements

Strong Python engineering skills on large codebases, including architecture, refactoring, typing, packaging, and performance profiling.
Experience with developer tooling, test infrastructure, or CI at scale.
Fluency with JIT compilation, code generation, or GPU computing stacks to reason about compilation caching, kernel dispatch, and startup costs.
Comfort debugging across multiple platforms (Windows, Linux, macOS) and architectures (x86, arm64), and GPU computing stacks (CUDA, AMD, Apple Metal).
Open-source experience, including triaging issues, reviewing external pull requests, and community communication.
Understanding of physics and robotics, with a working picture of how simulators like Genesis-World function and are used.
Opinions about API design.

Nice To Haves

Experience with other robotics simulators (Drake, MuJoCo, Newton, Isaac, RaiSim, Jiminy, Brax).
Experience with PyTorch.
Experience with other JIT compilers such as Triton, JAX, or Numba.
Experience with 3D tooling and interactive editors (Maya, Blender, Houdini).
Familiarity with serialization formats (USD, glTF, MJCF/URDF).
Experience with C++, CUDA, or Metal.
Developer relations or community work experience.

Responsibilities

Build the software engineering backbone of Genesis-World, ensuring reliability, maintainability, and frictionless use of the simulation codebase.
Make testing of a JIT-compiled engine scalable, addressing bottlenecks related to compile time and runtime.
Ensure any bug can be reproduced, including bit-exact simulation dumps and reloads across different machines and backends.
Develop functionality to turn any failing simulation run into a standalone reproduction script automatically.
Enhance the first-class support for automatic differentiation (autodiff) within Genesis-World, working on memory efficiency, performance, and maintainability.
Enable editing scenes at frozen time, allowing for instant iteration on scene layout without full build cycles.
Make the simulator extensible with first-class plugins for physics solvers and interactive viewer backends.
Integrate real actuation capabilities for motors and MIMO transmissions.
Decouple the simulation clock and timestep to allow independent environment stepping, adaptive timesteps, and individual solver advancement.
Implement event triggers based on physics, making scenes more dynamic.
Refactor legacy subsystems, strengthen typing, enforce consistent coding style, reduce Python overhead on hot paths, implement full scene serialization, and improve telemetry and replay.
Review issues and pull requests with the core team.
Answer questions from users and contributors.
Contribute to adoption efforts.