Pay
$166K–$237K is typical
The middle half of what similar US postings state. This one doesn’t list pay.
From 1,505 postings in the last year. Stated ranges, not salaries paid.
San Francisco · San Francisco, California, United States
Live, first seen by us 42 days agoLocation type: On-siteFrom the employer’s posting
Employment type: Full-timeFrom the employer’s posting
Department: Software EngineeringFrom the employer’s posting
Architect multi-tenant infrastructure for rate limiting, isolation, versioning and observability
Turn partner data into a product flow from upload through validation and evaluation
Build and operate low-latency inference systems for Pi models controlling robots
Turn recurring partner friction into better APIs, tooling, and documentation
Architect reliable, multi-tenant infrastructure around rate limiting, isolation, backpressure, versioning, observability, and SLOs.
Turn partner data into a first-class product experience, from initial upload through validation, processing, fine-tuning, and evaluation.
Build and operate low-latency inference systems that allow Pi models to control robots running in real-world environments.
Learn from how partners use the platform and turn recurring friction into better APIs, tooling, documentation, and abstractions.
Physical Intelligence is bringing general-purpose AI into the physical world.
In the employer’s words · Read in context · Company website ↗
These dates show when we found the posting. Check the employer’s website to confirm it is still accepting applications.
$166K–$237K is typical
The middle half of what similar US postings state. This one doesn’t list pay.
From 1,505 postings in the last year. Stated ranges, not salaries paid.
Day 42
Similar jobs usually come down between day 7 and day 77.
From 5,865 similar postings that came down. Coming down is not always a hire.
6 open jobs
Open jobs by the week we first saw them, last 12 weeks.
Hiring recordAdd your resume and we’ll mark which of these tools you already use, and point out anything in the posting that could rule you out.
- Enough familiarity with machine learning systems to deploy, serve, and debug models in production. You do not need to be an ML researcher. - Strong Python skills and the ability to work comfortably across infrastructure and product boundaries. - A high degree of ownership. You’re comfortable taking an ambiguous problem, building the first version end to end, and then designing the system so it no longer depends on you.
- Experience at an early-stage infrastructure, AI, robotics, or autonomous systems company. - Familiarity with our stack: Python, Postgres, ClickHouse, GCP, Kubernetes, Modal, React, and TypeScript. - Experience with security, authentication, authorization, or multi-tenant infrastructure.
Complete your application on jobs.ashbyhq.com. The employer’s form will show what is required.
Already applied? Track this application
Physical Intelligence is bringing general-purpose AI into the physical world. We are a group of engineers, scientists, roboticists, and company builders developing foundation models and learning algorithms to power the robots of today and the physically-actuated devices of the future.
We’ve made significant progress toward building models that can control many different robots across a wide range of tasks. Now, we’re building the platform that will make those models accessible to the broader world.
As an API Product Engineer, you’ll build the product surface that allows other companies to use Pi’s models much like developers use an LLM API today: bring their own data, fine-tune models, evaluate them, and run low-latency inference in their own environments.
You’ll own this experience end to end, turning capabilities that today require close collaboration with our team into a platform that can eventually support thousands—and ultimately millions—of robots.
Software Engineering at Pi builds the systems that scale research breakthroughs into reliable products in the physical world.
This role sits at the intersection of our AI, infrastructure, and external partners. You’ll work closely with researchers and engineers to turn rapidly evolving model capabilities into APIs and tools that external users, from hobbyists to large organizations, can actually build on.
Today, onboarding a new partner to Pi can involve significant hands-on work. The goal is to make that increasingly unnecessary: partners should be able to bring us data, fine-tune a model, evaluate it, and run inference through a product that is intuitive, reliable, and built to scale.
This is a software and systems role first. Robotics experience is not required, and we’re particularly excited about people who have built developer platforms, model APIs, or infrastructure products outside of robotics.
Build Pi’s model API end to end, including data ingestion, fine-tuning, evaluation, low-latency remote inference, partner-facing tools, and deployment integrations.
Design systems that can scale from a small number of deeply integrated partners to thousands of organizations and potentially millions of robots.
Architect reliable, multi-tenant infrastructure around rate limiting, isolation, backpressure, versioning, observability, and SLOs.
Turn partner data into a first-class product experience, from initial upload through validation, processing, fine-tuning, and evaluation.
Build and operate low-latency inference systems that allow Pi models to control robots running in real-world environments.
Work closely with researchers to turn new model capabilities into stable, usable product surfaces.
Learn from how partners use the platform and turn recurring friction into better APIs, tooling, documentation, and abstractions.
Write high-quality production code that integrates deeply with Pi’s existing infrastructure.
Define what the developer platform for general-purpose robotics should look like. There is no established playbook for this yet.
Strong software engineering fundamentals and experience building production systems.
Deep backend and systems experience across APIs, services, databases, caching, distributed systems, and infrastructure.
Experience building and scaling developer platforms. We’re especially interested in people who have shipped platforms for model fine-tuning, inference, or other compute-intensive workloads.
An understanding of the problems that emerge as systems scale: reliability, latency, multi-tenancy, versioning, observability, and operational complexity.
Enough familiarity with machine learning systems to deploy, serve, and debug models in production. You do not need to be an ML researcher.
Strong Python skills and the ability to work comfortably across infrastructure and product boundaries.
A high degree of ownership. You’re comfortable taking an ambiguous problem, building the first version end to end, and then designing the system so it no longer depends on you.
Experience building low-latency or real-time systems, including streaming, inference transport, WebSockets, QUIC, or similar technologies.
Experience building model-serving, inference, fine-tuning, or developer-platform infrastructure.
Experience at an early-stage infrastructure, AI, robotics, or autonomous systems company.
Familiarity with our stack: Python, Postgres, ClickHouse, GCP, Kubernetes, Modal, React, and TypeScript.
Experience with security, authentication, authorization, or multi-tenant infrastructure.
This is not a Customer Success role. You won’t own ongoing partner support or change management inside customer organizations.
This is not a Partner Roboticist or Applied Research role. You won’t be responsible for training policies for individual partners or debugging robots on site.
This is not a traditional ML infrastructure role focused primarily on training systems, compilers, or accelerator scheduling. You’ll work closely with those systems, but your focus is the product and platform through which the outside world uses Pi’s models.
Pursuant to the San Francisco Fair Chance Ordinance, we will consider for employment qualified applicants with arrest and conviction records.
New ML Engineer postings at the employers we follow are up 12% over the last 6 weeks compared with the 6 before.
Each bar is one week of new postings. The darkest bars are the last 6 weeks.
Counted from the hiring pages we’ve followed for at least five weeks, so employers we add don’t look like growth. Updated weekly.
We file this job under ML Engineer, by the work the posting describes.
ML engineers put machine learning models into products. They write the code that trains, ships, serves and monitors models so a recommendation or prediction works for real users.
Postings ask for strong Python and software engineering, a framework such as PyTorch, cloud services, and more and more often experience with large language models.
A data scientist often works out which model to build; an ML engineer makes it run reliably in production.
The government has no code for ML engineers, so these figures are for the closest group, computer and information research scientists.
Source: US Bureau of Labor Statistics, Occupational Employment and Wage Statistics (May 2025) and Employment Projections (2025 to 2035). Occupation code 15-1221.
Selected passages from the saved posting. Check the full description for conditions and exceptions.
No pay amount identified in the saved description.
San Francisco, California, United States
This is a software and systems role first. Robotics experience is not required, and we’re particularly excited about people who have built developer platforms, model APIs, or infrastructure products outside of robotics. - Build Pi’s model API end to end, including data ingestion, fine-tuning, evaluation, low-latency remote inference, partner-facing tools, and deployment integrations. - Design systems that can scale from a small number of deeply integrated partners to thousands of organizations and potentially millions of robots.
- This is not a Customer Success role. You won’t own ongoing partner support or change management inside customer organizations. - This is not a Partner Roboticist or Applied Research role. You won’t be responsible for training policies for individual partners or debugging robots on site. - This is not a traditional ML infrastructure role focused primarily on training systems, compilers, or accelerator scheduling. You’ll work closely with those systems, but your focus is the product and platform through which the outside world uses Pi’s models.
No clear work-authorization passage found. Eligibility is unconfirmed.