Principal AI Performance Modeling Architect
Role details
Job location
Tech stack
Job description
As a Principal Engineer, you will spearhead the next generation of AI infrastructure by defining GPU architecture specifications that enable massive model training at scale. Your expertise will drive 2-3x performance gains in both training and inference pipelines through innovative system design and optimization. You will champion the adoption of cutting-edge techniques across the engineering organization, from efficient attention mechanisms to advanced parallelization strategies. By establishing comprehensive best practices for distributed ML systems, you will create a framework that enables seamless scaling from single-GPU to thousand-GPU deployments. THE PERSON: You have a deep understanding of GPU microarchitecture, memory hierarchies, and their impact on large-scale ML workloads You are passionate about software engineering and possess leadership skills to drive sophisticated issues to resolution. You are able to communicate effectively and work optimally with different teams across AMD., * Lead performance modeling and optimization for multi-trillion parameter LLM training/inference including Dense, Mixture of Experts (MoE) with multiple modalities (text, vision, speech)
- Model/optimize novel parallelization strategies across tensor, pipeline, context, expert and data parallel dimensions
- Architect memory-efficient training systems utilizing techniques like structured pruning, quantization (MX formats), continuous batching/chunked prefill, speculative decoding
- Incorporate and extend SOTA models such as GPT-4, Reasoning models (Deepseek-R1), and multi-modal architectures
- Collaborate with internal and external stakeholders/ML researchers to disseminate results and iterate at rapid pace., AMD does not accept unsolicited resumes from headhunters, recruitment agencies, or fee-based recruitment services. AMD and its subsidiaries are equal opportunity, inclusive employers and will consider all applicants without regard to age, ancestry, color, marital status, medical condition, mental or physical disability, national origin, race, religion, political and/or third-party affiliation, sex, pregnancy, sexual orientation, gender identity, military or veteran status, or any other characteristic protected by law. We encourage applications from all qualified candidates and will accommodate applicants' needs under the respective laws throughout all stages of the recruitment and selection process. AMD may use Artificial Intelligence to help screen, assess or select applicants for this position. AMD's "Responsible AI Policy" is available here. This posting is for an existing vacancy.
Requirements
Do you have experience in System performance optimization?, Do you have a Bachelor's degree?, * Extensive and Senior experience optimizing large-scale ML systems and GPU architectures
- Deep expertise in CUDA programming, GPU memory hierarchies, and hardware-specific optimizations
- Proven track record architecting distributed training systems handling large scale systems
- Expert knowledge of transformer architectures, attention mechanisms, and model parallelism techniques
PREFERRED EXPERIENCE:
- PyTorch, CUDA, TensorRT, OpenAI Triton
- Distributed systems: Ray, Megatron-LM
- Performance analysis tools: NSight Compute, nvprof, PyTorch Profiler
- KV cache optimization, Flash Attention, Mixture of Experts
- High-speed networking: InfiniBand, RDMA, NVLink
ACADEMIC CREDENTIALS:
- Bachelors, MS/PhD in Computer Science/Engineering or equivalent industry experience