Experience Requirements:

• 7+ years of professional experience in C++ software development

• Focus on performance-critical applications

• Strong understanding of C++ templates and modern memory management

• Hands-on experience with SIMD programming (AVX-512, SSE)

• Experience in HPC, numerical computing, or ML inference optimization

Other Requirements:

• Experience with hardware accelerators (FPGAs, GPUs, or custom ASICs)

• Proficiency with profiling tools (perfetto, VTune, Valgrind)

• Knowledge of multi-threading and NUMA architectures

Responsibilities:

• Design and implement high-performance inference software for LLMs

• Develop and optimize C++-based libraries using SIMD and templates

• Optimize model execution via low-level optimizations and vectorization

• Apply NUMA-aware memory management to optimize access patterns

• Implement ML system-level optimizations like token streaming and batching