Elevate edge AI capabilities at Amazon Devices as a High-Performance ML Kernel Performance Engineer. Focus on crafting CUDA and Triton kernels for efficient model training and inference.
Join the AI Platform team, where you'll bridge hardware and software by optimizing performance at the GPU level. Your work will enhance compression algorithms, ensuring vast efficiency improvements as neural networks scale. Collaborate with scientists and engineers to democratize optimization and boost overall productivity across projects.
Key Responsibilities:
• Design and implement CUDA and Triton kernels for edge AI models
• Analyze kernel performance, resolving bottlenecks for faster training
• Optimize kernels through techniques like operator fusion and memory access
• Build tools for team members to test and profile kernel efficiency
• Extend the training kernels library with clean interfaces and CI
Requirements:
• 3+ years of professional software development experienc...