GPU Performance Engineer (AI)

TL;DR

GPU Performance Engineer (AI): Building and optimizing the foundational GPU stack for task-specific LLMs with an accent on latency-bound serving and throughput-bound training. Focus on writing high-performance kernels in CUDA/Triton and profiling complex workloads to deliver state-of-the-art performance for enterprise-scale deployments.

Location: Must be based in or be able to relocate to the Paris or New York office

Company

hirify.global is a well-funded startup developing a reinforcement learning platform to tune, evaluate, and serve specialized language models.

What you will do

• Build and maintain fast, robust GPU code for real-world AI applications.
• Develop high-quality software using CUDA, CUTLASS, or Triton.
• Profile and optimize dedicated GPU kernels across compute-bound and latency-bound regimes.
• Contribute to the product roadmap by identifying and implementing performance-enhancing trends.
• Collaborate within a distributed team using asynchronous communication.

Requirements

• M.Sc./Ph.D. in computer science or equivalent experience in software engineering.
• Strong background in systems programming and GPU optimization.
• Track record of writing high-performance kernels with state-of-the-art results.
• Demonstrated contributions to open-source projects like CUTLASS, Triton, or MLIR.
• Must be able to work on-site in Paris or New York (visa sponsorship available for relocation).

Culture & Benefits

• Comprehensive medical, dental, and vision insurance.
• 401(k) plan with 4% matching.
• Unlimited PTO policy with a 5-week annual minimum recommendation.
• Stipends for mental health, wellness, and personal development.
• Opportunity to shape early-stage research and product strategy.