GPU Kernel Engineer for High-Performance AI Inference

Ensimag Alumni; Paris; Sur place; EUR 50 000 - 70 000; Plein temps

Résumé du poste

Ensimag Alumni is seeking a talented engineer in Paris to work on GPU optimization for cutting-edge LLM inference at Kog. You'll engage in writing and profiling GPU kernels, optimizing computational paths, and contributing to a monokernel pipeline.

The ideal candidate has experience with CUDA or similar environments and a strong background from a top engineering school or relevant PhD. Join a dynamic team and push the boundaries of GPU engineering.

Qualifications

• Experience in writing GPU kernels with performance constraints.
• Understanding of hardware and GPU architecture.
• Familiarity with inline PTX or CDNA ISA.

Responsabilités

• Understand GPU internals and optimize computational sections.
• Contribute to a monokernel pipeline for LLM inference.
• Build profiling infrastructure for GPU programs.

Connaissances

GPU optimization

CUDA programming

PyTorch custom ops

Kernel optimization

Formation

Top engineering school or PhD with GPU work

Description du poste

Ensimag Alumni is seeking a talented engineer in Paris to work on GPU optimization for cutting-edge LLM inference at Kog. You'll engage in writing and profiling GPU kernels, optimizing computational paths, and contributing to a monokernel pipeline.

The ideal candidate has experience with CUDA or similar environments and a strong background from a top engineering school or relevant PhD. Join a dynamic team and push the boundaries of GPU engineering.