Founding ML Research Engineer (AI for Hardware Engineering)

A small startup is working to improve how well AI can handle real hardware engineering. It evaluates engineers at aerospace, defense, manufacturing and robotics firms, which yields expert-calibrated scoring criteria and recorded human workflows, and it wants a founding researcher to convert that material into training signal for models.

Your focus

• Design and build reinforcement learning environments plus evals aimed at agentic mechanical CAD and electronic design work
• Create the scoring graders, reward functions and checking mechanisms that rate engineering output dependably
• Package assessment data as datasets and environments usable by frontier AI labs in research and large training runs
• Partner with mechanical and electrical engineers who decide what counts as correct, and with research groups at labs that train models on the environments
• Conduct experiments and contribute to research and product direction

About you

• Three or more years in machine learning engineering, with senior-level initiative and ownership
• Experience with RL environments, whether you built and sold them or used them in post-training
• A hardware link, such as ML at a hardtech company, close work with EEs or MEs, or an engineering degree in those fields
• Knowledge of CAD, ECAD, 3D vision or similar hardware-adjacent ML
• Hands-on evaluation, grading, data pipeline and ML infrastructure work, and a history of training models rather than only building on top of them

Hybrid in the San Francisco Bay Area. Unable to consider candidates needing visa sponsorship or transfer.