Staff Machine Learning Engineer, Spot Autonomy

Boston Dynamics•Waltham, MA

3d•$155,000 - $210,000

About The Position

In this role as a Staff Machine Learning Engineer, you will have the opportunity to develop Spot’s next generation autonomy capabilities. You will research & integrate state-of-the-art approaches in the areas of perception, localization and navigation to ensure that our robots can navigate the world robustly and confidently. You will get to: Develop Spot’s next generation localization, mapping, perception and autonomy capabilities and deliver them to the product. Apply and develop novel ML-based approaches to solve complex challenges in semantically-aware navigation and localization. Shape the team's strategy for ML model architectures, datasets, and pipelines to achieve maximum model performance. Produce high-quality, performant code in C++ and Python. Design and execute rigorous experiments using both simulated and real-world robot data to ensure solutions not only achieve state-of-the-art performance but are also robust and computationally efficient in real-world deployments. As a part of the Spot Autonomy Team, you will closely collaborate with other skilled researchers & engineers who are passionate about Spot’s autonomy capabilities. Being embedded in the broader Spot R&D team, you will get a chance to further collaborate with other groups and experts from a wide variety of backgrounds. To succeed in this role, you should have the following skills and experience

Requirements

A Masters degree in Computer Science, Robotics or related field and 3+ years of professional experience
A solid understanding of traditional computer vision, robotics and navigation methods (SLAM) and their typical strengths & shortcomings
First-hand experience in Machine Learning and data driven approaches to visual perception problems.
A good understanding of recent ML approaches such as LLMs & ViTs.
Experience with Machine Learning frameworks (e.g. PyTorch)
Experienced in writing performant, well-structured, and testable C++ and Python code
Be a team player and good communicator, able to work well in a dynamic and collaborative environment
Have a passion for quality and autonomous robots

Nice To Haves

A PhD in Computer Science, Robotics or related field.
A strong background in SLAM / factor graphs (e.g. gtsam, g2o, Ceres), and ideally ML-enhanced navigation, such as semantic SLAM.
Experience with end-2-end semantic navigation approaches.
A track record of relevant publications or product deliverables.

Responsibilities

Develop Spot’s next generation localization, mapping, perception and autonomy capabilities and deliver them to the product.
Apply and develop novel ML-based approaches to solve complex challenges in semantically-aware navigation and localization.
Shape the team's strategy for ML model architectures, datasets, and pipelines to achieve maximum model performance.
Produce high-quality, performant code in C++ and Python.
Design and execute rigorous experiments using both simulated and real-world robot data to ensure solutions not only achieve state-of-the-art performance but are also robust and computationally efficient in real-world deployments.