Performance Engineer (ML)

Advanced Micro Devices, Inc•San Jose, CA

About The Position

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. SMTS SOFTWARE DEVELOPMENT ENGINEER THE ROLE: AMD is seeking an experienced ML Compiler Engineer to join our team and drive performance optimization for machine learning workloads that run on our datacenter GPU infrastructure. In this role, you will work at the intersection of compiler technology and ML, focusing on the LLVM compiler backend to generate highly efficient code for our GPU architectures. You will identify performance bottlenecks in ML workloads, conduct deep technical analysis, and implement compiler-level solutions that directly impact the efficiency and cost-effectiveness of our AI infrastructure. This position offers the opportunity to contribute to both our internal compiler toolchain and upstream open source projects, making your work visible to and benefiting the broader ML and compiler communities. THE PERSON: The ideal candidate is a seasoned compiler engineer with deep expertise in LLVM backend development and a proven track record of optimizing code generation for specialized hardware architectures. You are passionate about performance optimization and thrive on the challenge of extracting maximum efficiency from complex hardware. You have strong analytical skills that allow you to diagnose performance issues at the compiler level and devise creative solutions. As an active participant in the open source community, you understand the importance of collaborative development, code review processes, and the standards required for upstream contributions. You communicate technical concepts clearly, work effectively with cross-functional teams including hardware architects and ML researchers, and are driven by the impact of your optimizations on real-world AI workloads running at scale.

Requirements

Strong background in computer architecture and software engineering practices
Experience designing and working with compiler backends
Active contribution history to the open-source LLVM project
Demonstrated mastery of LLVM backend components, including LLVM IR, Machine IR (MIR), SelectionDAG, GlobalISel, Register Allocation, TableGen, and Instruction Scheduling
Effective communication and problem-solving skills
Master’s or Ph.D. in Computer Science, Computer Engineering, or a related field, with multiple years of professional experience in compiler development, with a significant focus on LLVM or other modern compiler backend
This role is not eligible for visa sponsorship.

Responsibilities

Identify and analyze performance bottlenecks in ML workloads at the compiler level, using profiling tools and performance analysis methodologies
Design and implement optimization passes in the LLVM backend to improve code generation quality for datacenter GPU architectures
Develop and enhance instruction selection, scheduling, and register allocation strategies specific to ML computation patterns
Collaborate with hardware architecture teams to ensure compiler optimizations align with and fully exploit GPU microarchitectural features
Contribute bug fixes, optimizations, and new features to upstream LLVM projects, following open source development practices
Maintain and improve internal LLVM-based compiler toolchains, ensuring compatibility with upstream changes while adding proprietary enhancements
Create comprehensive benchmarks and test suites to validate compiler optimizations and prevent performance regressions
Document compiler optimizations, architectural decisions, and best practices for both internal and external audiences
Work with ML framework teams to understand and optimize compilation of real-world models and workloads
Participate in code reviews, design discussions, and technical presentations to share knowledge across the organization