Real-Time SLAM & Spatial AI Research Engineer — 3D Perception

Aimlroles; Waltham (MA); On-site; USD 177,000 - 225,000; Full time

Benefits offered by this job

Medical

Dental

Vision

401(k)

Paid time off

Annual bonus

Job summary

Boston Dynamics is seeking an SLAM & Spatial AI Research Engineer to join the Atlas VLA Research team. You will design real-time SLAM, state estimation, and offline 3D reconstruction pipelines to give Atlas a grounded 3D understanding of the world and support large VLM/VLA training.

You will write robust C++ and Python code, collaborate with perception, robotics, ML, and system software teams, and test on state-of-the-art hardware. A PhD or equivalent experience is preferred.

Qualifications

• PhD or equivalent experience in robotics, computer vision, ML, CS, or related fields.
• Experience building and deploying SLAM, VO, or 3D reconstruction systems for robots or autonomous platforms.
• Strong background in real-time SLAM, state estimation, 3D reconstruction, and probabilistic estimation.

Responsibilities

• Design and implement real-time SLAM and perception-based state estimation for mobile humanoids or data collection devices.
• Build offline 3D reconstruction pipelines to generate geometric scene structure for model training and evaluation.
• Pioneer research integrating large VLA/VLM models with 3D spatial perception for language-grounded scene understanding.
• Bridge classical geometry with learned methods, choosing appropriate estimation strategies.
• Write high-quality C++ and Python code that fits into a large production codebase.

Skills

Real-time SLAM

Visual odometry

3D reconstruction

State estimation

C++

Python

Education

PhD in Robotics/CS/ML (or equivalent)

Tools

PyTorch

JAX

Job description

Boston Dynamics is seeking an SLAM & Spatial AI Research Engineer to join the Atlas VLA Research team. You will design real-time SLAM, state estimation, and offline 3D reconstruction pipelines to give Atlas a grounded 3D understanding of the world and support large VLM/VLA training.

You will write robust C++ and Python code, collaborate with perception, robotics, ML, and system software teams, and test on state-of-the-art hardware. A PhD or equivalent experience is preferred.