Simulation Engineer

Humble Robotics•San Francisco, CA

About The Position

We’re looking for a software engineer to join our simulation team. You’ll build the systems and tools that support closed-loop model testing, synthetic data generation, and sensor simulation. This work will work heavily with modern neural rendering approaches: Gaussian splatting, world models, and learned sensor models. We’re a small team early in the build. You’ll have real influence over both what we build and how we build it.

Requirements

BS, MS, or PhD in Computer Science, Electrical Engineering, Robotics, or a related field—or equivalent industry experience
Strong proficiency in Python (primary language of the simulation stack)
Experience working with simulation systems, ideally in robotics, autonomous vehicles, or other high-fidelity domains
Eligible to work in the United States

Nice To Haves

Experience with neural rendering or generative simulation (e.g., Gaussian splatting, NeRFs, world models, diffusion-based methods)
Depth in one or more areas of simulation: closed-loop testing, scenario design, log replay, or sensor simulation
Experience building benchmarks or evaluation frameworks used across a team
Experience developing synthetic data pipelines that improved downstream model performance
Familiarity with real-time robotic systems and their constraints on simulation fidelity
Comfortable working on a small team with high ownership and fast iteration cycles

Responsibilities

Build and maintain Python pipelines and tooling for closed-loop testing and synthetic data generation
Integrate modern neural rendering techniques into our sensor simulation stack (e.g., Gaussian splatting, NeRFs, world models)
Expand and maintain the scenario library and benchmark suite used to evaluate perception, planning, and control
Develop synthetic data generation workflows, including scenario and agent behavior generation, to support model training
Work closely with ML and autonomy engineers to measure and reduce sim-to-real gaps, ensuring simulation reflects real-world behavior