Research Scientist - Offline RL for Real-Robot Control

Brahma Consulting Group; San Francisco (CA); On-site; USD 150,000 - 230,000; Full time

Job summary

Brahma Consulting Group is recruiting a researcher to own offline RL research for a well-funded robotics startup in San Francisco. You will work with an existing researcher on problems already defined, with room to pursue your own directions.

Your work will run on real robots, not just in simulation. You will train robot policies from large mixed-quality datasets, explore value learning with sparse rewards, and develop model-based RL and world models for robot control.

Qualifications

• First-author RL papers at top ML venues.
• Research depth and taste over general engineering skills.
• Background in robotics or continuous control, or video/image generation with strong RL foundations.

Responsibilities

• Train robot policies from large, mixed-quality datasets, including video without action labels.
• Advance value learning for sparse rewards.
• Develop model-based RL and world models for robot control.
• Ensure RL holds up as data and model size grow.
• Run clean experiments with reproducible baselines.

Skills

Offline RL

Robotics

Continuous control

Video RL

Job description

Brahma Consulting Group is recruiting a researcher to own offline RL research for a well-funded robotics startup in San Francisco. You will work with an existing researcher on problems already defined, with room to pursue your own directions.

Your work will run on real robots, not just in simulation. You will train robot policies from large mixed-quality datasets, explore value learning with sparse rewards, and develop model-based RL and world models for robot control.