Research Engineer, Diffusion Rendering & 3D NeRFs

Inria; France; À distance; EUR 27 000 - 33 000; Plein temps

Avantages offerts par ce poste

Transport reimbursement

Annual leave 7 weeks + RTT

Teleworking possible

Equipment provided

Social events

Vocational training

Mutual insurance

Résumé du poste

Inria is seeking a Research Engineer for Machine Learning for and by Rendering to join the GRAPHDECO project. You will design and scale data pipelines, and implement GPU‑accelerated training to fine‑tune diffusion models integrated with neural scene representations.

The role combines software engineering with cutting‑edge graphics/ML research, offering exposure to physics‑based rendering and interactive 3D scene editing. Teleworking is possible with flexible hours.

Qualifications

• Master’s degree in Computer Science with focus on computer graphics and ML for visual computing.
• Computer Vision education is a plus.
• Proficiency in Python, PyTorch, C++, and CUDA for high‑performance deep learning and graphics workflows.
• Background in computer graphics (rendering, rasterization, ray tracing, path tracing, mitsuba3).
• Experience with training/fine tuning large models, diffusion models and transformers, automated data generation pipelines, multi‑GPU scaling.
• Experience with AI agentic coding frameworks for task automation.

Responsabilités

• Design, implement, and scale core computational and data infrastructure for the project.
• Define and deploy automated data collection, synthetic rendering, and validation pipelines.
• Implement GPU‑accelerated training workflows to fine‑tune diffusion models and integrate with NeRFs/3D representations.
• Build robust software abstractions and benchmarking frameworks for reproducible dynamic simulations.

Connaissances

Python

PyTorch

C++

CUDA

Rendering

Ray tracing

Path tracing

Mitsuba3

Diffusion models

Transformers

Data pipelines

Multi-GPU scaling

AI agentic coding

Formation

Master’s degree in Computer Science (Graphics/ML)

PhD preferred

Outils

Python

PyTorch

C++

CUDA

Description du poste

Inria is seeking a Research Engineer for Machine Learning for and by Rendering to join the GRAPHDECO project. You will design and scale data pipelines, and implement GPU‑accelerated training to fine‑tune diffusion models integrated with neural scene representations.

The role combines software engineering with cutting‑edge graphics/ML research, offering exposure to physics‑based rendering and interactive 3D scene editing. Teleworking is possible with flexible hours.