AI Framework Engineer

WHAT YOU DO AT AMD CHANGES EVERYTHING   At AMD, our mission is to build great products that accelerate next-generation computing experiences—from AI and data centers, to PCs, gaming and embedded systems. Grounded in a culture of innovation and collaboration, we believe real progress comes from bold ideas, human ingenuity and a shared passion to create something extraordinary. When you join AMD, you’ll discover the real differentiator is our culture. We push the limits of innovation to solve the world’s most important challenges—striving for execution excellence, while being direct, humble, collaborative, and inclusive of diverse perspectives. Join us as we shape the future of AI and beyond.   Together, we advance your career.

As a core member of the team, you will play a pivotal role in  optimizing  and developing deep learning frameworks for AMD GPUs. Your experience will be critical in enhancing GPU kernels, deep learning models, and training/inference performance across multi-GPU and multi-node systems. You will engage with both internal GPU library teams and open-source maintainers to ensure seamless integration of optimizations,  utilizing   cutting-edge  compiler technologies and advanced engineering principles to drive continuous improvement.   THE PERSON:     Skilled  engineer with strong technical and analytical  expertise  in C++ development within Linux environments. The ideal candidate will thrive in both collaborative team settings and independent work, with the ability to define goals, manage development efforts, and deliver high-quality solutions. Strong problem-solving skills, a proactive approach, and a keen understanding of software engineering best practices are essential.

Deep Learning & LLM Framework Optimization:  Optimize major DL/LLM frameworks (TensorFlow,  PyTorch ,  vLLM ,  SGLang ) for AMD GPUs and contribute improvements upstream.   GPU Kernel & Operator Optimization:  Develop and tune GPU kernels and performance-critical operators to maximize throughput and minimize latency.   Model & Architecture Optimization:  Adapt and optimize LLM architectures (e.g., Llama, Qwen, DeepSeek) and apply advanced techniques like  FlashAttention ,  PagedAttention , and quantization.   End-to-End Performance Engineering:  Perform comprehensive profiling to identify bottlenecks and implement system, memory, and communication optimizations across multi-GPU and multi-node setups.   Compiler & Pipeline Acceleration:  Leverage advanced compiler technologies and graph compilers to enhance the full deep learning and inference pipeline.   Research & Advanced Techniques:  Prototype and integrate emerging optimization methods such as speculative decoding and weight-only quantization into production systems.   Cross-Team & Open-Source Collaboration:  Collaborate with internal GPU library teams and open-source maintainers to align improvements and ensure seamless upstream integration.   Software Engineering Excellence:  Apply robust engineering practices to deliver maintainable, reliable, and production-quality performance optimizations.   MANDATORY EXPERIENCE:     Inference Frameworks, Model Architectures & Optimization Expertise: Strong practical experience with vLLM  or  SGLang , mastery of modern LLMs (e.g., DeepSeek, Qwen), strong theoretical grounding in Transformer/Attention/ MoE /KV Cache, and hands-on application of advanced inference optimizations such as  FlashAttention ,  PagedAttention , continuous batching, and quantization (INT8/INT4/GPTQ/AWQ).   End-to-End LLM Performance Engineering:  Demonstrated ability to profile, diagnose, and optimize  compute , memory, and communication bottlenecks across multi-GPU and multi-node environments.   High-Performance Computing: Expert e xperience running and optimizing large-scale workloads on heterogeneous clusters with a focus on efficiency, reliability, and