Senior Software Engineer, Spatial Intelligence and Foundation Models

NVIDIA is seeking a Senior Software Engineer to join the Isaac Spatial Intelligence team within the Isaac Engineering organisation. You will work on geometric and semantic understanding, and reasoning for robots, building perception systems that turn raw sensor data into actionable world understanding, shaping the future of physical AI!

Responsibilities:

• Design, implement, and deploy novel algorithms for spatial understanding, working on problems such as SLAM, structure-from-motion, optical flow, scene flow, and object reconstruction, and training VLMs on various spatial reasoning skills.
• Develop robust perception, mapping, and reasoning systems that run on robots, in simulation, and at scale in data pipelines for training foundation models.
• Advance the state of the art in geometric computer vision, combining classical multi-view geometry and optimisation with modern deep learning approaches.
• Train and evaluate vision-language models on skills relevant to robotics.
• Collaborate with the Cosmos and GR00T teams on perception and spatial understanding for foundation models, and work with other Isaac teams, including Sim/Lab, Platform, and SQA.
• Contribute to integrating, validating, and releasing applied research efforts in collaboration with research teams and on top of NVIDIA's advanced robotics platforms.
• Foster a culture of innovation and collaboration, supporting deliverables such as prototypes, open-source software contributions, patents, and publications.
• Work cross-functionally with product, hardware, and software teams to translate engineering work into impactful products.

Requirements:

• PhD or Master's degree in Computer Science, Robotics, or a related field (or equivalent experience).
• 8+ years of experience working on computer vision, robotics, or deep learning technologies.
• Strong foundation in 3D geometric computer vision: multi-view geometry, visual odometry/SLAM, structure-from-motion, or dense correspondence (optical flow, scene flow, stereo).
• Strong hands-on programming skills in Python and/or C++; experience with Deep Learning frameworks (PyTorch, JAX, TensorFlow).
• Experience training and evaluating deep learning models for perception tasks; familiarity with vision-language models is a strong plus.
• Strong skills in agentic tool use , proficiency working with AI coding agents and agentic workflows (e.g., Claude Code, Cursor, Codex) to accelerate development, testing, and experimentation.
• Excellent communication, organisational, and interpersonal skills.

Nice to Have:

• Contributions to widely used SLAM, SfM, or 3D reconstruction systems (open-source or shipped products).
• Experience with large-scale model training on GPU clusters, including VLMs or other foundation models.
• Publications at top computer vision or robotics venues (CVPR, ICCV, ECCV, ICRA, IROS, RSS, CoRL).
• Hands-on experience with simulation-based training and evaluation, sim-to-real and real-to-sim transfer.

You will also be eligible for equity and benefits.