Deep Learning Engineer Intern – 3D Reconstruction

Sourced from internexxus.com

PlusAI is seeking a Deep Learning Engineer Intern to work on 3D reconstruction for autonomous driving. You will develop pipelines to create high-fidelity digital twins using neural rendering techniques like 3D Gaussian Splatting and NeRFs. This role involves collaborating with simulation and perception teams to validate autonomous truck systems.

• Construct Digital Twins: Develop and optimize pipelines to reconstruct high-fidelity, multi-modal 3D representations of key driving routes using recorded vehicle data including camera, LiDAR, IMU, and GPS.
• Implement Next-Gen Tech: Apply state-of-the-art neural rendering and view-synthesis techniques, such as 3D Gaussian Splatting (3DGS) and NeRFs, to handle challenging real-world lighting, weather, and dynamic objects.
• Drive Validation & Verification (V&V): Integrate reconstructed environments into our perception testing framework to stress-test object detection, tracking, and edge-case scenarios.
• Evaluate Multi-Modal Accuracy: Ensure alignment and geometric consistency between different sensor modalities within the reconstructed digital twin.
• Collaborate & Scale: Work closely with the Simulation and Perception teams to turn research prototypes into robust, scalable tooling that directly impacts our production deployment timeline.