Michael G. Email & Phone Number

I am a computer engineer with broad experience spanning Compilers, Parallel and Accelerated Computing, ASIC Design and Machine Learning. I am currently at NVIDIA working on scaling LLM training as part of the end to end training performance group.Previously I was the compiler lead and one of the first employees at EnCharge AI where I helped bootstrap the SW engineering org and build ML compilers for a novel in-memory computing technology. At Qualcomm Research I worked on optimizing AI/ML applications for Snapdragon powered devices and developed new accelerator architectures for low power on device inference. During my second tenure I was part of the AI100 compiler team solving challenging compute and memory scheduling problems.I spent some time earlier in my career at NVIDIA working on various projects in the Compute Architecture group for high performance training, specifically compilers, kernels (CUTLASS) and architecture for deep learning. I also spent some time at Waymo optimizing deep learning models for self-driving cars and trucks as well as working on various TensorFlow/XLA integration and deployment challenges.HW/SW experience includes modeling for design space exploration, verification and debugging, software optimization and codesign, ASIC design, embedded software/firmware, and custom compiler development. I primarily work in C++, Python and assembly code and when the need arises I can write RTL.ML experience includes CNNs, RNNs, transformers, mixture of experts, quantization and compression and various frameworks including TensorFlow, PyTorch and Jax.In the ML compiler space I have experience with several libraries and projects including LLVM, TVM, Glow, XLA and MLIR.

Listed skills include C++, Software Engineering, Python, C, and 5 others.