Spandan Tiwari

Spandan Tiwari#

Spandan

Spandan Tiwari is a Director of Engineering, AI Software at AMD. He leads a group focusing on algorithms research and software engineering for AI model optimization (quantization, sparsity, pruning, fusion) and delivering support for these models/techniques in various inference stacks for AMD devices (e.g. MI GPUs, Ryzen NPU, and EPYC CPUs). Prior to AMD, he was Engineering Manager at Meta in the PyTorch group, leading a team developing parts of the inference stack Meta’s custom AI accelerator. Before that, he was at Microsoft where he was one of the founding team members of the ONNX/ONNXRuntime standard and ecosystem. His research interests include model compression algorithms, HW/SW co-design, and computer vision. He received his PhD from the University of Missouri - Rolla in Electrical Engineering with a focus on computer vision.

Posts by Spandan Tiwari

https://rocm.blogs.amd.com/software-tools-optimization/mxfp4-online-rotation/README.html
https://rocm.blogs.amd.com/artificial-intelligence/hipblaslt_offline_tuning/README.html
https://rocm.blogs.amd.com/software-tools-optimization/mxfp4-mxfp6-quantization/README.html
https://rocm.blogs.amd.com/artificial-intelligence/mlperf-inference-v5.1/README.html
https://rocm.blogs.amd.com/artificial-intelligence/quick-reduce/README.html