Staff, ML Research Scientist

Boston Dynamics•Waltham, MA

About The Position

As a Staff ML Research Scientist on the Machine Learning Safety R&D Team, you will join a small cross-functional group developing machine learning models that will enable our robots to operate safely around people. Every day you will help research, design, and build innovative machine learning models and algorithms to run on our robots. Your work will pave the way for our Embodied AI. In this role you will chart a path by combining the best of state of the art ML architectures with real-world safe robotics challenges, ultimately creating novel solutions to one of the most important problems in robotics. If you are creative, thrive in a small team environment, and passionate about using machine learning to further a world where humans and robots truly work together - come join us!

Requirements

5+ years of experience applying Deep Learning (ML) to scalable real-world problems in computer vision or LLMs.
Experience in the full modeling cycle from research to deployment of modern Deep Learning architectures such as Transformers, VLMs/VLAs, and Deep Reinforcement Learning.
Knowledge of state of the art work in related areas of computer vision, world models, and Deep Reinforcement Learning.
Experience with the full lifecycle of Deep Learning development, including network design, data management, training, evaluation, hyperparameter search, deployment/productionalization, and online validation.
Strong communication skills, including ability to author technical documentation and deliver presentations on technical topics.
History of working in small, interdisciplinary teams.
Eligible to work in the United States.

Nice To Haves

Publications are a plus.

Responsibilities

Research and build novel model architectures which allow our robots to operate safely around people.
Research, build, validate, and deploy ML models to detect hazards, humans, and other environmental features.
Develop datasets, metrics, and validation plans for ML model research.
Work closely with a small team to design and prototype new deep learning based perception and behavior models which create safety features for our robots.