Robotics Software Engineer - World Models

About the Company

This role is with a stealth-mode robotics startup backed by a16z with a $120M seed round, founded by DeepMind pioneers who built the algorithms that made general robot intelligence possible. The company is building world models for humanoid robotics — the AI brain that lets robots understand and act in the physical world right out of the box. Further details about the company will be shared in later stages of the hiring process.

The Opportunity

The company is looking for a Robotics Software Engineer to work across the core world model stack — building the software systems that enable humanoid robots to perceive, reason, and act in novel environments without requiring task-specific programming.

This is a generalist robotics engineering role, not a narrow specialist position. You will work across perception, planning, control, and data infrastructure as the product demands, contributing wherever the highest-leverage engineering work is. Strong software fundamentals are the baseline — domain expertise in robotics and ML is what separates the right candidates. The best fits will come from teams at NVIDIA, Google, or leading robotics labs where they shipped real systems to real robots.

What You’ll Do

• • Build and improve the software systems powering humanoid robot world models across perception, planning, and control


• • Develop and maintain data pipelines that collect, process, and prepare real-world robot experience for model training


• • Implement and optimize inference infrastructure for deploying world models across a fleet of 100+ robots


• • Write production-grade code that runs reliably on deployed hardware and scales as the fleet grows


• • Collaborate closely with research and hardware teams to close the loop between simulation, training, and physical deployment

You Should Have

• • Strong software engineering fundamentals — algorithms, data structures, system design, and clean production code


• • Experience working on robotics software systems in production, not just research prototypes


• • Proficiency in Python and C++ for robotics and systems development


• • Familiarity with robot software stacks including ROS or equivalent frameworks


• • Comfort working across ambiguous problem spaces where the right abstraction is still being defined

Nice to Have

• • Experience at NVIDIA, Google, or a leading robotics or autonomous systems team


• • Background in world model research, imitation learning, or reinforcement learning for robotics


• • Experience with large-scale data pipelines or model deployment for physical robot fleets


• • Familiarity with simulation environments such as Newton or MuJoCo


• • Prior work on humanoid, cobot, or autonomous vehicle platforms