Core Simulation Engine Engineer

November 4, 2026

Hybrid

Senior

London, United Kingdom

Indeed

Also worth knowing

Mandatory office attendance

Description

An open-source, general-purpose multi-physics simulation platform for physical AI with an in-house real-time renderer and a cross-platform JIT compiler. The Core Simulation Engine Engineer joins as a core maintainer, building the software engineering backbone of the platform: infrastructure, tooling, APIs and architecture. The role centers on scalable testing of a JIT-compiled engine, reproducible debugging, first-class autodiff, plugin extensibility and performance across GPU backends.

Responsibilities

Serve as core maintainer building infrastructure, tooling, APIs and architecture around the physics algorithms
Make testing of the JIT-compiled engine scalable despite bad compile-to-runtime time ratios
Make bugs reproducible with bit-exact dump and reload plus automatic standalone reproduction scripts
Make autodiff first-class by reducing memory use and improving maintainability with the compiler team
Enable instant scene editing at frozen time without a full build cycle or engine recompile
Add first-class plugins for physics solvers and interactive viewer backends including the in-house renderer
Abstract motors and MIMO transmissions cleanly at the user-to-physics-engine interface
Free the clock to sub-scene level so environments and solvers can step independently
Add physics-triggered events so scenes react to presses, triggers and contacts
Sharpen the core through refactoring legacy subsystems, stronger typing and consistent coding style
Reduce Python overhead on hot paths and add full scene serialization, telemetry and replay
Review issues and pull requests and answer questions from users and contributors

Requirements

Strong Python engineering on large codebases including architecture, refactoring, typing and packaging
Performance profiling experience with py-spy, Nsight, Xcode Instruments and RenderDoc
Developer tooling, test infrastructure or CI experience at scale
Solid understanding of JIT compilation, code generation or GPU computing stacks including compilation caching and kernel dispatch
Comfort debugging across Windows, Linux and macOS on x86 and arm64 with CUDA, AMD and Apple Metal
Open-source reflexes: triaging issues, reviewing external pull requests and communicating with the community
Comfort with physics and robotics concepts relevant to a simulation engine
Bonus: experience with robotics simulators such as Drake, MuJoCo, Isaac, Newton or Brax
Bonus: familiarity with PyTorch, JIT compilers like Triton, JAX or Numba, and 3D tooling like Maya or Houdini