Computer Vision Engineer Intern

About the role

DevExplore is a Physical AI firm building virtual driver software for autonomous trucking. This internship position involves creating 3D and 4D digital twins and neural rendering for our simulation and perception departments.

Key facts

What you'll do

• Build pipelines to create high-fidelity 3D representations using sensor data from LiDAR, cameras, GPS, and IMU.
• Use 3D Gaussian Splatting and NeRFs to model dynamic objects, weather, and lighting.
• Generate 3D and 4D scene reconstructions to simulate driving environments and produce synthetic sensor data.
• Incorporate these environments into testing frameworks to evaluate perception systems and edge cases.
• Verify geometric consistency and sensor alignment within digital twins.
• Transition research prototypes into scalable tools for production use.

Requirements

• Currently enrolled in an MS or PhD program in Statistics, Mathematics, Electrical Engineering, Computer Science, or a related field.
• Solid grasp of machine learning principles.
• Experience building deep learning models using JAX, TensorFlow, or PyTorch.

Nice to have

• Experience with 3D computer vision.
• Background in motion tracking, semantic segmentation, or object detection.
• Publications in venues such as SIGGRAPH, AAAI, NeurIPS, ECCV, ICCV, ICLR, or CVPR.

Skills & tools

• PyTorch, TensorFlow, JAX
• 3D Gaussian Splatting (3DGS), NeRFs
• LiDAR, camera, IMU, GPS sensor data processing

Practical notes

• Hourly pay ranges from $19 to $65, depending on academic level, experience, and location.
• Benefits include a 401(k) plan, free catered lunches, and unlimited snacks and drinks.