AI Software Lead PyTorch & CUDA Runtime (Next-Gen Accelerator)

Role Overview

We are looking for a Software Lead (8+ years experience) to own the runtime and neural network (NN) layer of a next-generation AI accelerator platform. This role focuses on designing, optimizing, and implementing NN operators and developing new ops using CUDA/custom runtime APIs to deliver high-performance execution on custom AI hardware.

Key Responsibilities

• Design and optimize NN operators for performance-critical workloads
• Develop new NN ops using CUDA/custom runtime APIs
• Drive runtime-level optimizations across compute, memory, and scheduling
• Own runtime NN layer interfaces and execution model
• Implement and optimize operator fusion (e.g., matmul + bias + LayerNorm) for efficient hardware utilization
• Identify and resolve performance bottlenecks across the stack
• Collaborate with compiler, PyTorch framework, and low-level SW teams

Impact

• Own how efficiently AI workloads execute on the platform
• Drive performance, scalability, and hardware utilization through optimized runtime and NN ops design

Qualifications

Required Qualifications

• 8+ years in systems software / runtime / performance engineering
• Strong experience with:
  • PyTorch / TensorFlow / JAX or similar frameworks
  • NN operator/kernel development and optimization
  • operator fusion and graph-level optimizations
• Hands-on expertise in:
  • C/C++ and CUDA (or similar low-level programming)
  • runtime systems and execution engines
• Strong understanding of:
  • memory hierarchy, data movement, and parallel execution

Preferred Qualifications

• Experience with GPU/NPU or custom AI accelerators
• Familiarity with XLA / MLIR / compiler-runtime interaction
• Experience optimizing LLM or large-scale DL workloads