Robotics AI API Platform Engineer

Physical Intelligence; San Francisco (CA); On-site; USD 150,000 - 210,000; Full time

Job summary

Physical Intelligence is hiring an API Product Engineer to build the developer-facing surface that enables customers to use Pi’s models—bring their data, fine-tune, evaluate, and run low-latency inference in their environments.

You’ll own this experience end to end, turning capabilities into a scalable platform that can support thousands of partners and millions of robots, with strong collaboration across AI, infrastructure, and external users.

Qualifications

• Strong software engineering fundamentals and production systems experience.
• Experience building developer platforms, model APIs, or infrastructure products.
• Understanding of reliability, latency, multi-tenancy, observability, and operational complexity.
• Familiarity with ML systems to deploy and debug models in production.
• Strong Python skills across infrastructure and product boundaries.

Responsibilities

• Build Pi’s model API end to end: data ingestion, fine-tuning, evaluation, and low-latency inference.
• Design scalable systems for thousands of partners and millions of robots.
• Architect multi-tenant infrastructure with rate limiting, isolation, and observability.
• Turn partner data into a first-class product experience from upload to evaluation.
• Build and operate low-latency inference systems for real-world robots.
• Collaborate with researchers to turn new model capabilities into stable surfaces.
• Learn from partner usage to improve APIs, tooling, and documentation.
• Write production-grade code deeply integrated with Pi’s infra.
• Define the developer platform vision where there is no established playbook.

Skills

Software engineering

Backend systems

Developer platforms

Python

ML systems

Ownership

System scalability

Tools

Postgres

ClickHouse

Kubernetes

GCP

Modal

React

TypeScript

Job description

Physical Intelligence is hiring an API Product Engineer to build the developer-facing surface that enables customers to use Pi’s models—bring their data, fine-tune, evaluate, and run low-latency inference in their environments.

You’ll own this experience end to end, turning capabilities into a scalable platform that can support thousands of partners and millions of robots, with strong collaboration across AI, infrastructure, and external users.