Deep Learning Performance Architect, CUTLASS DSL Testing

Are you excited about building world-class quality systems for advanced GPU software? Do you enjoy combining automation, product validation, and code analysis to support fast-moving compiler and kernel innovation? We are seeking a strong test engineer to develop the NVIDIA CUTLASS DSL testing framework, shape product test strategy, and ensure end-to-end code quality across the MLIR-based compilation pipeline. In this role, you will drive automated testing, and regression detection to make sure every code change is validated for correctness, and the product is ready for shipping at any time.

What you'll be doing:

• Develop and evolve the NVIDIA CUTLASS DSL testing framework for next-generation GPU software
• Define, refine, and execute robust product test strategies for shipping to the open-source community
• Ensure end-to-end code quality across the MLIR-based compilation pipeline and related functional coverage infrastructure
• Build automated testing, code coverage measurement, and regression detection workflows at scale
• Partner with multiple teams to make sure every operator change meets a high bar for correctness, quality, and performance

What we need to see:

• MS, PhD, or equivalent experience in Computer Science, Software Engineering, or a related field
• 3+ years of relevant work experience
• Excellent Python and scripting skills
• Strong experience developing and using test tools, with a solid understanding of software testing best practices
• Hands-on experience with automated testing in GPU environments, including correctness testing, code coverage improvements, and regression detection
• Strong communication skills and proven ability to collaborate effectively across teams

Ways to stand out from the crowd:

• Familiarity with common AI agent technologies and applications
• Experience in quality assurance of open-source products

Seen 12 days ago · NVIDIA postings close after a median of 43 days.

Original posting on NVIDIA's site ↗

Posting text belongs to the employer. Removal requests: contact us.