Robotics Software - Perception ML Engineer

The role

You will develop and deploy perception models for manipulation and navigation in dynamic outdoor environments. You will work on 3D scene understanding, detection, segmentation, pose, and world/video models using multimodal data.

What you will do

• Develop and deploy state-of-the-art perception models for manipulation and navigation problems.
• Drive the technical approach with a focus on real-world deployments.
• Research and develop techniques for 3D scene understanding, detection, segmentation, pose and world/video models.
• Develop robust metrics to establish performance of large vision models.
• Work with multimodal data including cameras, LIDAR, and tactile.
• Develop and improve real/synthetic data pipelines.
• Build scalable training and validation infrastructure.
• Collaborate with other teams in the company.

Required qualifications

• M.S/Ph.D degree in robotics, vision, computer science, mechanical engineering, electrical engineering or other engineering disciplines (or equivalent experience).
• 4+ years of hands-on experience developing perception solutions for robotics applications like manipulation and navigation.
• Experience working on one or more of detection, segmentation, pose estimation, tracking, 3D reconstruction, world models and VLMs.
• Experience working with synthetic data pipelines and robotics simulation infrastructure (e.g. Isaac Sim, Gazebo, Blender, Unity).
• Extensive experience with Python and common ML frameworks like PyTorch.
• Comfortable taking ownership of tasks with light supervision.
• Excellent problem-solving skills.
• Legally authorized to work in the United States.

About the company

Gritt is an intelligent system that combines robotics and AI to build infrastructure. The founding team includes experts from Carnegie Mellon, Stanford, and MIT.