Rendering Engineer, Light Fields

Department: Engineering & Research

Location: Paris

Employment Type: FullTime

About the role

Genesis AI is building Nyx, a renderer that delivers offline quality at realtime performance, powered by next generation capture and rendering technology built on light fields and Gaussian splatting. We are looking for a principal level rendering engineer to own our light field pipelines end to end, from capture and training through Gaussian splats or other suitable representations all the way to color accurate output on real displays. Nyx is a hybrid renderer optimized for robotics scenarios. This is an ownership role, not a pure research position, where you will be responsible for the product itself rather than just the papers behind it.

What you'll do

• Design and implement light field generation and reconstruction pipelines

• Build and optimize Gaussian splat generation, training, compression, and real-time rendering

• Own color science end to end: camera characterization, color spaces, tone mapping, HDR output, and display calibration

• Move research into production, turning papers (yours or others') into robust, performant code

• Collaborate with capture, ML, and platform teams to define the pipeline architecture

What we're looking for

• PhD in computer graphics, computational imaging, computer vision, or a related field (or an equivalent research track record)

• Publications at venues such as SIGGRAPH, SIGGRAPH Asia, CVPR, ICCV, or EGSR

• Hands-on experience with radiance fields, 3D Gaussian splatting, or light field rendering beyond toy implementations

• Solid grounding in color science: colorimetry, camera ISP pipelines, and ACES or similar managed workflows

• Strong C++ and GPU programming (CUDA, compute shaders), plus Python for research tooling

Nice to have

• Industry experience shipping capture, VFX, or real-time rendering products

• Differentiable rendering, novel view synthesis, or neural compression experience

• Familiarity with HDR standards (PQ, HLG) and display characterization