Physical AI Engineer

Physical AI Engineer – Simulation & Synthetic Data

Build the Future of Robotics with Physical AI

We're building real‑world Physical AI systems—where learning agents interact with physical machines. This role is tailored for an AI Robotics Engineer with a strong reinforcement learning (RL) mindset, someone who wants to train, evaluate, and deploy intelligent behaviors that emerge through interaction, not just perception.

You'll work hands‑on with NVIDIA Omniverse, physics‑based simulation, synthetic data, and foundation models to train agents that learn to act, adapt, and generalize in complex environments. This is a builder role: fast iteration, scalable training, and direct transfer from simulation to physical robots.

What You'll Do

• Build high‑fidelity, physics‑based simulation environments in NVIDIA Omniverse / Isaac Sim for training and evaluating robotic agents.
• Design and run reinforcement learning and imitation learning pipelines using simulation‑generated and synthetic data.
• Train and tune policies for control, planning, and decision‑making, emphasizing robustness and sim‑to‑real performance.
• Integrate foundation models (GPT‑class, Claude / Opus, multimodal models) to support reasoning, task decomposition, or human‑in‑the‑loop learning.
• Drive simulation‑to‑real transfer, ensuring learned behaviors perform reliably on physical robotic systems.
• Collaborate with robotics, controls, and perception engineers to deploy learning‑based systems end to end.

Requirements

• 5–8+ years of experience in robotics, reinforcement learning, simulation, or applied machine learning.
• Strong hands‑on experience with reinforcement learning, imitation learning, or learning‑based control.
• Proven experience building and training agents in simulation environments (synthetic data, domain randomization, curriculum learning).
• Experience working in the NVIDIA ecosystem (Omniverse, Isaac Sim, USD pipelines strongly preferred).
• Experience applying or integrating foundation models (e.g., GPT‑based or Claude / Opus‑style models) into robotics or decision‑making workflows.
• Strong Python skills; ability to build and scale training pipelines.
• Builder mindset with a track record of moving learning systems from experiment to deployment.

Nice to Have

• Model‑free or model‑based RL at scale
• Sim‑to‑real transfer techniques (domain randomization, system ID, hybrid control)
• Physics engines, real‑time systems, or robotics middleware
• Experience operationalizing learned policies on physical robots

Work Environment

• Remote / hybrid (US‑based)
• Occasional domestic and global travel
• Flexible working hours aligned to experimentation and training cycles

Compensation

Annual compensation offered will be based on several variables including geographic location, work experience, education, and skills and achievements, and will be mutually agreed upon at the time of offer. Hiring Salary Range (USD): $ 134,500 – $168,200

Why Join

You'll work on learning‑driven robotics systems where intelligence is earned through interaction with the world—not just inferred from data. If you're excited by agents that learn, fail, adapt, and ultimately succeed in the real world, this role is built for you.