Skip to main content

Medra

AI Product Engineer

San FranciscofulltimemidAdded today

About this role

Join a startup building Physical AI systems that autonomously conduct scientific research. You'll develop AI-driven workflows and ship end-to-end features that help scientists accelerate discovery using robotics and intelligent agents.

What you'll do

  • Design and implement agentic workflows for multi-step scientific tasks with planning, execution, and verification
  • Build customer-facing AI product features from prototype through production deployment
  • Develop production-grade APIs, data pipelines, and distributed services for AI-driven systems
  • Contribute to product roadmap decisions informed by AI capabilities and customer needs
  • Collaborate with team to iterate rapidly on features under ambiguity

What they're looking for

  • Python and/or TypeScript
  • AI/ML product development and workflow design
  • Production systems architecture (APIs, data pipelines, distributed services)
  • Product thinking and scoping under uncertainty
  • Robotics or physical systems integration
  • Full-stack software development
  • Agentic AI frameworks and LLM integration

Benefits

  • Significant equity ownership
  • Medical and dental insurance
  • 401k
  • Unlimited PTO
  • Weekday dinners
  • Relocation assistance available
Apply with Autofill

Opens the application — the Jobs AI extension fills it for you. Set up autofill

Opens the official application on the employer’s site. No login required.

Medra

Medra develops a Physical AI Scientist platform that uses robotic automation to innovate life science research and laboratory workflows. The company is hiring Mechanical Engineers to design and test hardware for robotic systems that enhance laboratory protocols.

View all jobs at Medra

Likely interview questions

  • Describe a time you shipped an AI-driven product feature end-to-end—what were the key challenges in moving from prototype to production?
  • How would you design an agentic workflow to autonomously execute and verify a multi-step scientific experiment?