Lead AI Compiler & Kernel Optimization for NPUs

microTECH Global LTD; Warszawa; On-site; PLN 260,000 - 380,000; Full time

Job summary

microTECH Global LTD is seeking a world-class compiler and performance optimization expert to join our deep learning infrastructure team in Warsaw. You will own the Triton compiler and kernel optimization framework, driving AI workloads on NPUs with advanced techniques in compiler theory and system performance engineering.

You will lead the design and implementation of state-of-the-art kernels, optimize memory patterns and scheduling, and integrate Triton with PyTorch and other backends to

Qualifications

• Master's degree or above in Computer Architecture, Compiler Theory, High Performance Computing or related field.
• PhD preferred for research track and advanced optimization tasks.
• 5+ years of experience in NPU/GPU programming, operator optimization, or compiler development.

Responsibilities

• Lead design and development of the Triton compiler and performance optimization framework for high-performance NPUs.
• Implement state-of-the-art Triton kernels (Attention, MatMul, LayerNorm, Conv, Softmax) with top-tier efficiency.
• Optimize memory access patterns, scheduling and cache usage to maximize SM occupancy and throughput.
• Integrate Triton with PyTorch, XLA, and runtime stacks to enable end-to-end performance gains.
• Research auto-tuning, kernel fusion, and operator scheduling to scale performance across models.
• Mentor team members in Triton kernel development and establish standard performance-analysis processes.
• Stay current with MLIR, TVM, Hidet, Cutlass and related compiler tech to push performance boundaries.
• Perform performance modeling and workload fingerprinting for large models to guide optimizations.

Skills

NPU/GPU programming

Triton

Kernel development

Performance optimization

Auto-tuning

Kernel fusion

Operator scheduling

Memory optimization

Profiling tools

Education

Master's degree in Computer Architecture, Compiler Theory, HPC

PhD preferred

Tools

Triton

LLVM

TVM

MLIR

Cutlass

PyTorch

Job description

microTECH Global LTD is seeking a world-class compiler and performance optimization expert to join our deep learning infrastructure team in Warsaw. You will own the Triton compiler and kernel optimization framework, driving AI workloads on NPUs with advanced techniques in compiler theory and system performance engineering.

You will lead the design and implementation of state-of-the-art kernels, optimize memory patterns and scheduling, and integrate Triton with PyTorch and other backends to