Machine Learning Research Engineer
- Compensation
- Not disclosed
- Location
- San Francisco
- Work setup
- Onsite
- Seniority
- Not specified
- Employment
- Full time
- Experience
- Not specified
- Category
- Ml / AI
- Posted
Skills
Must have
- Strong machine learning fundamentals and practical research engineering skills.
- Experience implementing, training, or evaluating models.
- Ability to design useful experiments and reason carefully about results.
- Strong programming skills and comfort working across research and production code.
- Interest in physical robots and willingness to test ideas against real-world behavior.
Nice to have
- Experience with robotics learning, computer vision, multimodal models, or control.
- Experience with simulation, robot datasets, or real-world evaluation.
- Experience optimizing inference on constrained hardware.
- Open-source work, research projects, or shipped ML systems that demonstrate your approach.
Full description
Machine Learning Research Engineer
San Francisco · Onsite · Full-time
About Bracket Bot
We believe planet earth can thrive and be alive. Robots should manufacture goods, clean roadways, tidy our rooms, build more homes, cook healthy food, and lower the cost of living. We aren’t quite there yet, robots are too expensive, too difficult to program, fragile, and unfriendly. Bracket Bot is changing that.
We believe that robots are currently in the ENIAC era and we are bringing them to the Macbook era. We believe in high performance engineering teams where everyone you work with is what you imagined Bell Labs employees to be like. Intelligent, scrappy, dedicated, kind.
At Bracket Bot, we have a few objectives. We are working to make the ideal role model of a company, one where everyone is at the top of their field, inventing new ways to solve the problems we face across mechanical, electrical, software, and machine learning. What does a company that creates entirely new ways that teams work together look like? A company that invents new materials for lighter weight actuators and structures? A company that creates and publishes new inverse kinematics algorithms for robots? A company that is thinking radically differently, and executing on another level.
In every field there are small teams who did something their peers had decided was impossible. The 10-second barrier in sprinting, or the eight engineers who left Motorola and made the 6502 for 10x lower cost than other chips at the time making the personal computer possible. SpaceX has landed the same rocket 36 times. Bracket Bot believes actuated, alive machines can be glorious, affordable, easy to use, and ubiquitous, and we are pulling that future forward.
Our products are a mirror of what we think about the world, on engineering & research, and what we think should be introduced into the world. In history, there are simply products that should exist, and only once created, are obvious in hindsight. Bracket Bot’s mission is to introduce as many of these products into the world as possible. We believe there are hundreds of these in the space of robotics. Across physical robot hardware, software and operating system; the infrastructure and distributed systems they connect to; the AI inference they run, both local and cloud, and so much more.
Our first product, Bracket Bot 1, exemplifies what a small team of highly talented people can do. At Bracket Bot, we genuinely believe 100x engineers exist, and a room full of them can create things that dozens of other companies have failed to do. We’ve brought to market a robot that is 10x lower cost than any other robot in its category, while out performing them on manipulation, autonomous navigation, speech interaction, and developer UX.
We’re backed by Fifty Years (Scott Phoenix), BoxGroup, SV Angel, Betaworks, Pace Capital, angels including Nat Friedman, Logan Kilpatrick (Google AI), Guillermo Rauch (Vercel) and others from Tesla AI, Physical Intelligence, and Markforged.