Senior Simulation Engineer I/ii, Robotics

Lila Sciences Lila Sciences · AI Frontier · Alewife, Cambridge, MA · Robotics

This role focuses on building the foundational simulation and robotic infrastructure for an AI Science Factory, involving scene authoring, automated data generation pipelines, physics and sensor fidelity, and sim-to-real transfer workflows. The engineer will work with NVIDIA Isaac Sim, Omniverse Kit, and USD to create reusable simulation libraries and tooling.

What you'd actually do

  1. Design and maintain modular, reusable simulation scene libraries using well-structured USD (Universal Scene Description), including proper use of references, payloads, variants, layers, and composition arcs.
  2. Build Python-based tooling and automation around the Omniverse Kit API (`omni.usd`, `omni.kit`, `omni.replicator`) to enable programmatic scene construction, editing, and batch processing, including headless Isaac Sim workflows for large-scale automated simulation runs and synthetic data generation.
  3. Contribute to internal simulation SDKs, APIs, and developer tools that allow other engineers to build on top of the simulation platform.
  4. Prototype and evaluate new simulation techniques, including physics calibration, sensor modeling, environment generation, and GPU-accelerated motion planning (e.g., cuRobo) for inverse kinematics and collision-aware trajectory generation.
  5. Design and implement domain randomization strategies (appearance, physics, lighting, sensor noise) using NVIDIA Replicator, and build tooling to evaluate and benchmark sim-to-real performance across configurations.

Skills

Required

  • 3–5 years of hands-on experience in simulation engineering, robotics software, or a closely related field.
  • Proficiency with NVIDIA Isaac Sim and the Omniverse Kit framework, including Python scripting via the Omniverse API (`omni.usd`, `omni.kit.app`, `omni.physx`).
  • Solid understanding of USD (Universal Scene Description), including authoring modular scenes with references, payloads, variants, sublayers, and composition arcs.
  • Familiarity with robot kinematics concepts (forward/inverse kinematics, motion planning, joint control) and implementing them programmatically in a simulation environment.
  • Strong Python skills; comfort writing clean, maintainable, production-quality code.
  • Experience with physics simulation concepts (rigid body dynamics, contact/friction modeling, articulations).

Nice to have

  • Experience running Isaac Sim in headless mode for automated, script-driven simulation workflows, and with distributed or cloud-based simulation workloads.
  • Experience with cuRobo or similar GPU-accelerated motion generation libraries (CUDA-based trajectory optimization, batched IK), and motion planning frameworks like MoveIt2 or OMPL.
  • Familiarity with NVIDIA Replicator for synthetic data generation and domain randomization, OmniGraph for node-based compute, and Isaac Lab for RL-based training environments.
  • Experience with ROS/ROS2 and sensor simulation (cameras, depth, LiDAR), including real-to-sim workflows (photogrammetry, NeRF, CAD asset pipelines).
  • C++ experience relevant to simulation extensions or performance-critical tooling.

What the JD emphasized

  • simulation engineering
  • robotics software
  • NVIDIA Isaac Sim
  • Omniverse Kit framework
  • USD (Universal Scene Description)
  • robot kinematics concepts
  • Python skills
  • physics simulation concepts
  • headless mode
  • distributed or cloud-based simulation workloads
  • cuRobo
  • motion planning frameworks
  • NVIDIA Replicator
  • OmniGraph
  • Isaac Lab
  • ROS/ROS2
  • sensor simulation
  • real-to-sim workflows
  • C++ experience

Other signals

  • simulation
  • robotics
  • synthetic data generation
  • sim-to-real transfer