10 hours ago
Responsibilities
- Conduct external benchmarking and research to identify GPU kernel performance improvements for AMD machine learning products.
- Analyze and optimize GPU kernels and identify optimization paths across the AI software stack.
- Partner with customers and hyperscalers to understand workload requirements and deliver architectural and software optimizations.
- Collaborate with hardware architecture, compiler, and framework teams to influence future GPU silicon features.
- Improve ROCm applications, libraries, tools, and hardware through collaboration with AMD experts, partners, and customers.
- Define and drive end-to-end software optimization strategy for high-performance customer solutions.
Requirements
- Strong understanding of modern model architectures, including Transformer, Attention, and KV Cache.
- Experience with optimization techniques such as quantization, speculative decoding, and FlashAttention.
- Excellent GPU architecture knowledge, performance analysis, algorithm development, and machine learning experience.
- Proven history optimizing GPU kernels with advanced algorithms across the AI software stack.
- Ability to drive cross-functional initiatives in fast-paced and ambiguous environments.
- Bachelor’s, master’s, or PhD degree in Computer Science, Computer Engineering, or equivalent.
Benefits
- Hybrid role located in Austin, Texas.
- This role is not eligible for visa sponsorship.
- AMD benefits are offered as described in the company’s benefits overview.
Categories
About AMD
We care deeply about transforming lives with AMD technology to enrich our industry, our communities, and the world. Our mission is to build great products that accelerate next-generation computing experiences – the building blocks for the data center, artificial intelligence, PCs, gaming and embedded. Underpinning our mission is the AMD culture. We push the limits of innovation to solve the world’s most important challenges. We strive for execution excellence while being direct, humble, collaborative, and inclusive of diverse perspectives. AMD together we advance_