Senior Machine Learning Engineer

DeepRec.ai; San Francisco (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

DeepRec.ai is seeking senior engineers to advance reconstruction technology at the core of its photorealistic driving simulation platform. You'll work on Gaussian Splatting, NeRF-based reconstruction, and multi-sensor rendering to push the state of the art in autonomous driving simulation.

The role emphasizes turning research into production code, with opportunities to collaborate with customers and shape benchmarks that quantify fidelity and the sim-to-real gap.

Qualifications

• 5+ years building and shipping ML or computer vision systems.
• Hands-on with 3DGS and/or NeRFs.
• Strong grounding in multi-view geometry, camera models and differentiable rendering.

Responsibilities

• Advance Gaussian Splatting and NeRF-based reconstruction for large dynamic driving scenes.
• Work on dynamic actor reconstruction, scene editing and novel view synthesis.
• Handle multi-sensor rendering (camera and LiDAR).
• Improve feed-forward reconstruction to reduce per-scene optimisation cost at scale.
• Build benchmarks for reconstruction fidelity and the sim-to-real gap.
• Collaborate directly with customers to solve real problems.

Skills

ML/CV systems

NeRFs

Multi-view geometry

Python programming

C++/CUDA

Tools

PyTorch

CUDA

CV frameworks

Job description

I'm partnering with a well-funded autonomy and simulation company whose software is used by most of the world's major car manufacturers and by defence customers.

The team builds photorealistic simulation from real-world driving data, and they're hiring senior engineers to improve the reconstruction technology at its core.

What you'll work on:

• Advancing Gaussian Splatting and NeRF-based reconstruction for large, dynamic driving scenes
• Dynamic actor reconstruction, scene editing and novel view synthesis
• Multi-sensor rendering (camera and LiDAR)
• Feed-forward reconstruction to cut per-scene optimisation cost at scale
• Building benchmarks for reconstruction fidelity and the sim-to-real gap
• Working directly with customers to solve real problems

What they're looking for:

• 5+ years building and shipping ML or computer vision systems
• Hands-on experience with 3DGS and/or NeRFs
• Strong grounding in multi-view geometry, camera models and differentiable rendering
• Python and PyTorch, plus C++ and/or CUDA
• A track record of turning research into production code
• Bonus: feed-forward/generalisable splatting, SLAM/SfM, graphics, AV or robotics data, and publications at CVPR, ECCV, NeurIPS, SIGGRAPH or similar.