Applied AI Engineer

Intel Intel · Semiconductors · Bangalore, India

This role focuses on designing and developing software features for AI frameworks, specifically optimizing them for Intel's AI accelerators and GPUs. The engineer will work on ML kernel development, enhance training and inference capabilities, and contribute to open-source AI communities like PyTorch, Tensorflow, and JAX.

What you'd actually do

  1. Design and develop SW features for AI frameworks - both HW-agnostic and HW-aware, especially in ML kernel development.
  2. Enhance and extend the Deep learning training, and Inference capabilities in the Software stack.
  3. Identifying optimization opportunities in the software stack to enhance performance of Deep learning workloads
  4. Participate in discussions with Open-source community, involve in development, adopting upstream and Upstream software.

Skills

Required

  • Advanced C++ (C++ 14/17)
  • Intermediate Python
  • Parallel programming
  • Developing machine learning kernels (GEMM, Convolution, Flash attention)
  • PyTorch, Tensorflow or JAX
  • Deep Learning models/LLMs for Vision / NLP
  • Debugging complex issues in multi-layered SW systems
  • Understanding of SW integration in large open-source frameworks
  • Computer architecture
  • HW-SW optimization techniques

Nice to have

  • developing and integrating CUTLASS or Triton based kernels in Large language models (LLMs)
  • compiler algorithms for heterogeneous system
  • Fuser optimizations

What the JD emphasized

  • frameworks
  • optimization
  • kernels
  • training
  • Inference

Other signals

  • Optimizing AI frameworks for Intel's hardware
  • Developing and optimizing ML kernels
  • Enhancing training and inference capabilities