Principal Software Engineer, Kernels

The role

You will develop, enhance, and maintain software kernels for next-generation AI hardware. You will work with compiler experts and hardware experts to productize the software stack for an AI compute engine.

What you will do

• Develop, enhance, and maintain software kernels for next-generation AI hardware
• Map computational graphs generated by AI frameworks to the underlying architecture
• Work with a team of compiler experts to build out the compiler infrastructure
• Work closely with ML, systems, mixed signal, DSP, and CPU experts
• Build and scale software deliverables in a tight development window

Required qualifications

• MS in computer engineering, math, physics, or a related degree with 10+ years of industry experience or a PhD in computer engineering, math, physics, or a related degree with 5+ years of industry expe
• Strong grasp of computer architecture, data structures, system software, and machine learning fundamentals
• Proficient in C/C++ and Python development in Linux environments and using standard development tools
• Experience implementing algorithms in high-level languages such as C/C++ and Python
• Experience implementing algorithms for specialized hardware such as FPGAs, DSPs, GPUs, and AI accelerators using libraries such as CUDA, etc.
• Experience in implementing operators commonly used in ML workloads—GEMMs, Convolutions, BLAS, SIMD operators for operations like softmax, layer normalization, pooling, etc.
• Experience with development for embedded SIMD vector processors such as Tensilica
• Self-motivated team player with a strong sense of ownership and leadership

Nice to have

• Prior startup, small team, or incubation experience
• Experience with ML frameworks such as TensorFlow and/or PyTorch
• Experience working with ML compilers and algorithms, such as MLIR, LLVM, TVM, Glow, etc.
• Experience with a deep learning framework (such as PyTorch or TensorFlow) and ML models for CV, NLP, or recommendation
• Work experience at a cloud provider or AI compute/subsystem company

About the company

d-Matrix has pioneered a memory-centric architecture that brings compute directly into memory. The team works on chips, software, and rack-scale systems to tackle AI infrastructure challenges.