Member of Technical Staff, Kernel Engineering

About the Company

Inferact's mission is to grow vLLM as the world's AI inference engine and accelerate AI progress by making inference cheaper and faster. Founded by the creators and core maintainers of vLLM, we sit at the intersection of models and hardware.

Responsibilities

• Write kernels and low-level optimizations to maximize performance on modern accelerators
• Work with hardware vendors to integrate new chips with vLLM
• Extract maximum performance from various hardware generations, from NVIDIA GPUs to emerging silicon
• Optimize code to run efficiently across hundreds of accelerator types

Requirements

• Bachelor's degree in computer science, engineering, or equivalent experience
• Deep experience writing CUDA kernels or equivalent (CuTeDSL, Triton, TileLang, Pallas)
• Strong understanding of GPU architecture including memory hierarchy, warp scheduling, tiling, and tensor cores
• Proficiency in C++ and Python for high-performance code
• Experience with profiling tools such as Nsight or rocprof
• Obsession with benchmarks and performance optimization

Preferred Qualifications

• Experience with ML-specific kernel optimization like FlashAttention or fused kernels
• Knowledge of quantization techniques including INT8, FP8, and mixed-precision
• Familiarity with multiple accelerator platforms such as NVIDIA, AMD, TPU, or Intel
• Experience with compiler technologies like LLVM, MLIR, or XLA
• Contributions to vLLM, open-source GPU/ML systems, or compiler optimization projects