Senior GPU Performance Engineer

Senior GPU Performance Engineer

Own performance analysis and optimization strategy across GPUs and accelerators, and turn your findings into automated compiler optimizations.

Hybrid, Full-time

Daisytuner is building the software layer for the next generation of computing. We make complex software run efficiently on any processor, from CPUs and GPUs to novel accelerators, using a self-learning compiler and cloud-scale optimization infrastructure.

Our team brings together researchers and engineers from RWTH Aachen, TU Munich, TU Darmstadt, and ETH Zurich to tackle some of the hardest problems in systems and infrastructure software. If you want to work on deeply technical challenges with real-world impact, join us and help shape the future of compute.

The Role

As Senior GPU Performance Engineer, you own performance analysis and optimization strategy across GPUs and other accelerators. You find bottlenecks in real-world workloads, investigate promising optimization techniques and build prototypes that show how code should ideally run on different hardware architectures.

You work closely with compiler engineers to turn these findings into automated compiler optimizations, performance models and tuning strategies. You also evaluate competing technologies and emerging hardware to find the most valuable optimization opportunities and shape the compiler's technical roadmap.

What you will be doing

• Analyze performance-critical workloads, especially machine learning and LLM applications, to find bottlenecks and optimization opportunities.
• Investigate how workloads interact with GPU architectures, including memory hierarchies, occupancy, instruction throughput, synchronization, launch configuration and execution patterns.
• Benchmark, profile and hand-tune GPU kernels, and develop optimization strategies with code transformations and prototypes to validate their performance impact.
• Work with compiler engineers to turn validated optimizations into work packages and milestones for automated compiler transformations, heuristics and tuning strategies.
• Prioritize optimization opportunities and turn performance findings into concrete goals and recommendations for the compiler roadmap.
• Evaluate competing compiler technologies, GPU programming frameworks, new GPU architectures and hardware features to find opportunities for Daisytuner.
• Build and maintain benchmark suites, evaluation methods and regression tests to track optimization quality across workloads and hardware architectures.

What you are bringing

• Master's degree in Computer Science, Electrical Engineering, Mathematics or a related technical field, or equivalent experience.
• 3+ years of GPU programming experience, including in a lead or ownership role.
• Proven optimization results, such as measurable speedups, shipped features or publications.
• Strong C/C++ and CUDA skills, with the ability to implement and prototype optimized GPU kernels.
• Experience analyzing and optimizing machine learning, HPC or other performance-critical workloads.
• Deep understanding of GPU memory hierarchies, shared memory, caches, register usage, occupancy, warp scheduling and instruction-level performance.
• Ability to independently investigate low-level performance bottlenecks, evaluate alternative implementations and reason about their performance impact.
• Hands-on experience with profiling tools such as Nsight Compute, Nsight Systems, CUPTI or rocProfiler.
• Ability to design systematic and reproducible performance evaluations.
• Strong Python skills for benchmarking, automation and experiments, and comfort with Linux and Git-based workflows.
• Ability to communicate findings clearly in English, turn them into actionable recommendations for compiler and research engineers, and guide technical direction without formal authority through discussions, design reviews and mentoring.
• Understanding of how performance improvements translate into customer value and product priorities.

Nice to have

• Experience across multiple GPU vendors and architectures (NVIDIA, AMD, Intel).
• Experience with GPU programming frameworks and libraries such as CUTLASS, Triton, cuBLAS, cuDNN, ROCm or SYCL.
• Familiarity with compiler infrastructures such as LLVM, MLIR or similar systems.
• Experience with kernel fusion, code generation or compiler optimization.
• Knowledge of transformer architectures, LLM inference/training or other modern ML workloads.
• Experience evaluating competing compiler technologies, performance optimization products or hardware-specific software stacks.
• Familiarity with performance modeling and automated optimization techniques.

What we offer

• A small, highly technical team where your work directly shapes our core technology.
• Real ownership of our GPU optimization strategy, with direct influence on the compiler and production workloads.
• Work on modern GPUs and accelerators across different hardware architectures.
• The opportunity to work at the intersection of compilers, machine learning and high-performance computing.
• Competitive compensation and potential equity participation.

We welcome applications from people of all backgrounds and do not discriminate on the basis of gender, ethnic origin, nationality, religion or belief, disability, age or sexual identity.