Pay
$166K–$237K is typical
The middle half of what similar US postings state. This one doesn’t list pay.
From 1,537 postings in the last year. Stated ranges, not salaries paid.
San Francisco · San Francisco, California, United States
Actively hiring, first seen by us 1 day agoLocation type: On-siteFrom the employer’s posting
Employment type: Full-timeFrom the employer’s posting
Department: EngineeringFrom the employer’s posting
Connect AI models to robot observations and actions
Build the tools, skills, and controllers models rely on
Develop evaluations in simulation and on the robot fleet
Decide when a model should act directly, coordinate policies, or write control code
Diagnose failures and fix them in observations, tools, policies, or orchestration
The work includes connecting models to robot observations and actions
building the tools, skills, and controllers they rely on
developing evaluations in simulation and on our fleet
when a frontier model should act directly, when it should coordinate faster learned policies or classical controllers, and when it should write its own tools or control code
When something fails, you'll work out whether it calls for a better observation, a different tool, a change in orchestration, a better policy or controller, or a fix elsewhere in the system.
Will you now or in the future require visa sponsorship to work in the United States?From the employer’s application form
Have deep experience in robotics across areas such as control, perception, planning, and learned policies
have made robots do useful work on real hardware
Have designed and shipped robotic or AI systems that people use, with clear evidence of what made them better
Are excellent at designing experiments and interpreting results
Can build reliable software across the stack, from model APIs and inference to robot software and testing infrastructure
Actively hiringFirst seen by us 1 day ago, still listed, and not a repost of an earlier posting.
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 1,537 postings in the last year. Stated ranges, not salaries paid.
Day 1
Early. Similar jobs usually come down between day 7 and day 76.
From 6,389 similar postings that came down. Coming down is not always a hire.
1 open job
Open jobs by the week we first saw them, last 12 weeks.
Hiring recordAdd your resume to compare the role’s scope, tools and requirements with your experience.
Complete your application on jobs.ashbyhq.com. The employer’s form will show what is required.
Already applied? Track this application
Pantograph is building general purpose intelligent robots that can safely and autonomously accomplish useful work in the real world. We believe the path to capable robots runs through general intelligence rather than narrow, robot-specific skills, and that there is an enormous amount of work worth doing in the physical world that would improve quality of life for people. As AI models make it easier to build software around our needs, we hope robots can empower people to shape the physical world.
Across research, hardware, and software, we focus on finding simple ideas that work, executing them extremely well, and scaling them up. We value clear thinking, careful execution, and systems we can understand, trust, and improve.
We're looking for a robotics ML engineer to design and build the systems that turn capable AI models into robots doing useful work. You'll bring a deep background in robotics and use it to decide how frontier models, learned policies, and classical controllers fit together. We see frontier models as an increasingly powerful foundation for general robot behavior, and want someone who can combine their capabilities with the methods robotics already does well.
The work includes connecting models to robot observations and actions, building the tools, skills, and controllers they rely on, and developing evaluations in simulation and on our fleet. You'll investigate when a frontier model should act directly, when it should coordinate faster learned policies or classical controllers, and when it should write its own tools or control code. You'll weigh the tradeoffs in capability, reliability, latency, safety, and how easily we can understand and supervise the system, with real deployments guiding your decisions.
You'll work closely with our research, robot software, and product teams, building a tight loop between experiments and experience on real hardware. We're looking for someone who combines excellent experimental judgment with a strong intuition for both robots and models, and enjoys building things to test their ideas. When something fails, you'll work out whether it calls for a better observation, a different tool, a change in orchestration, a better policy or controller, or a fix elsewhere in the system.
You might be a good fit if you:
Have deep experience in robotics across areas such as control, perception, planning, and learned policies, and have made robots do useful work on real hardware
Use simulation to develop and test systems, and understand where it diverges from reality
Have designed and shipped robotic or AI systems that people use, with clear evidence of what made them better
Are excellent at designing experiments and interpreting results: separating competing explanations, identifying confounders, and knowing what the evidence does and doesn't establish
Have a strong intuition for how models behave: what they understand, what confuses them, and how their behavior changes with the tools, context, and feedback they receive
Enjoy digging into failures and understanding how systems behave beyond aggregate metrics
Are pragmatic about methods and care more about what works on real tasks than what's newest or most complex
Can build reliable software across the stack, from model APIs and inference to robot software and testing infrastructure
Communicate clearly and enjoy working across research and engineering
Use AI tools extremely effectively
Bonus if you've worked with:
Agent harnesses, tool use, memory, or long-horizon tasks
Systems where models write and execute code or develop their own tools
Fine-tuning vision-language-action models or other large robot policies
Teleoperation, robot data collection, or learning from fleet data
Evaluation design, failure analysis, or safety testing
Inference optimization, including latency, throughput, and cost
We care much more about what you've built and learned than credentials. We're a small, fast-moving team working together in person in San Francisco. If this sounds exciting to you, we want to hear from you.
New ML Engineer postings at the employers we follow are down 4% 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
Working pattern and location restrictions need checking in the full posting.
No clear work-authorization passage found. Eligibility is unconfirmed.