Robotics ML Research Engineer – Real-World Robot AI

Bracket Bot Inc.; San Francisco (CA); On-site; USD 120,000 - 160,000; Full time

Job summary

Bracket Bot Inc. in San Francisco is seeking a Machine Learning Research Engineer to develop, evaluate, and integrate models for physical robots. You will work across research experiments, data, model evaluation, and inference to turn promising ideas into working robot behavior.

You’ll collaborate with robotics and software engineers, test on real hardware, and use observed failures to guide the next experiment, shaping how ML becomes part of the product at Bracket Bot.

Qualifications

• Strong ML 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.

Responsibilities

• Develop and evaluate machine learning approaches for robot perception and behavior.
• Build data and evaluation workflows that make model performance and failure modes clear.
• Run experiments, compare approaches, and communicate results and tradeoffs.
• Integrate models into robot software and test them on physical hardware.
• Improve inference performance and reliability within practical compute and latency constraints.
• Work with robotics and software engineers to turn research results into useful product capabilities.

Skills

ML fundamentals

Research engineering

Programming

Experiment design

Tools

Python

Job description

Bracket Bot Inc. in San Francisco is seeking a Machine Learning Research Engineer to develop, evaluate, and integrate models for physical robots. You will work across research experiments, data, model evaluation, and inference to turn promising ideas into working robot behavior.

You’ll collaborate with robotics and software engineers, test on real hardware, and use observed failures to guide the next experiment, shaping how ML becomes part of the product at Bracket Bot.