Senior GenAI Kernel Performance Engineer

Databricks; California (MO); On-site; USD 191,000 - 233,000; Full time

Job summary

Databricks is seeking a staff software engineer focusing on GenAI performance and kernel development. You will own high-performance GPU kernels for the GenAI inference stack, optimize for various hardware backends, and mentor teammates in kernel-level performance engineering.

You will collaborate with ML researchers, systems engineers, and product teams to push inference performance at scale, including profiling, optimization, and integration into production systems.

Qualifications

• Hands-on kernel tuning for ML workloads (CUDA, Triton, OpenCL, LLVM IR).
• Deep knowledge of GPU architectures and memory hierarchy.
• Experience with advanced optimization: tiling, fusion, auto-tuning.
• Familiarity with ML kernel libraries (cuBLAS, cuDNN, CUTLASS, oneDNN).
• Strong debugging and profiling skills (Nsight, NVProf, perf, vtune).
• Experience integrating optimized kernels into real-world ML inference systems.
• Track record of shipping high-performance production software.

Responsibilities

• Design, implement, benchmark, and maintain core compute kernels for GPU backends.
• Drive kernel-level performance improvements: vectorization, tiling, fusion, auto-tuning.
• Integrate kernel optimizations with higher-level ML systems.
• Build profiling and verification tooling for correctness and performance.
• Lead investigations of inference bottlenecks (memory bandwidth, cache).
• Define reusable kernel abstractions and cross-backend portability.
• Influence architecture decisions for memory layout and scheduling.
• Mentor engineers and perform code reviews.
• Collaborate with infra and ML teams to roll out optimizations.

Skills

Kernel tuning

CUDA

Triton

OpenCL

LLVM IR

GPU architecture

Profiling

ML inference

Distributed systems

Leadership

Education

BS/MS/PhD in Computer Science

Tools

Nsight

NVProf

perf

vtune

cuBLAS

cuDNN

CUTLASS

oneDNN

Job description

Databricks is seeking a staff software engineer focusing on GenAI performance and kernel development. You will own high-performance GPU kernels for the GenAI inference stack, optimize for various hardware backends, and mentor teammates in kernel-level performance engineering.

You will collaborate with ML researchers, systems engineers, and product teams to push inference performance at scale, including profiling, optimization, and integration into production systems.