Principal Software Engineer, Kernels

About the job

Responsibilities

• Develop, enhance, and maintain software kernels for next-generation AI hardware.
• Productize the software stack for our AI compute engine.
• Map computational graphs from AI frameworks to the underlying hardware architecture.
• Optimize and make trade-offs in hardware-software co-design.
• Build and scale software deliverables within tight development windows.
• Collaborate with compiler experts to build compiler infrastructure.
• Work closely with software (ML and systems) and hardware (mixed signal, DSP, CPU) experts.

Requirements

• MS in computer engineering, math, physics, or related with 10+ years of industry experience, OR PhD with 5+ years of industry experience.
• Strong understanding of computer architecture, data structures, system software, and machine learning fundamentals.
• Proficient in C/C++ and Python development within Linux environments.
• Experience implementing algorithms in high-level languages (C/C++, Python).
• Experience implementing algorithms for specialized hardware (FPGAs, DSPs, GPUs, AI accelerators) using libraries like CUDA.
• Experience implementing common ML workload operators (GEMMs, Convolutions, BLAS, SIMD operators for softmax, layer normalization, pooling).
• Experience with development for embedded SIMD vector processors (e.g., Tensilica).
• Self-motivated team player with strong ownership and leadership skills.
• Experience with ML frameworks such as TensorFlow and/or PyTorch (preferred).
• Experience with ML compilers and algorithms (MLIR, LLVM, TVM, Glow) (preferred).
• Experience with a deep learning framework and ML models for CV, NLP, or recommendation (preferred).
• Work experience at a cloud provider or AI compute/subsystem company (preferred).