Summary
Ramyad Hadidi is a Senior Staff ML Computer Architect with 11 years of experience bridging hardware and software to accelerate ML workloads, currently building low-TCO LLM inference datacenter accelerators with digital in-memory-compute at d-Matrix. He holds a PhD from Georgia Tech and an h-index of 22, reflecting impactful research in processing-in/near-memory, GPU/CPU architecture, and distributed ML for resource-constrained edge devices. Ramyad has led LLM workload analysis, SoC-level roofline and transaction modeling, and novel HW–SW mapping techniques for attention and GEMM to minimize memory footprint across SRAM and HBM. His background spans applied product optimization on edge platforms (Jetson, Snapdragon, ARM Ethos), prototyping on FPGA, and pioneering in-sensor and in-storage computing for CIS and SSDs. Colleagues rely on him for technically rigorous pathfinding, competitive analysis, and translating research into deployable accelerator designs. An uncommon thread across his work is combining deep theoretical insight with hands-on prototyping to close the loop from algorithm to silicon.
11 years of coding experience
10 years of employment as a software developer
Astrophysics and Astronomy International Olympiad Physics, Astrophysics and Astronomy International Olympiad Physics at Young Scholars Club
Doctor of Philosophy (PhD) Computer Science, Doctor of Philosophy (PhD) Computer Science at Georgia Institute of Technology
Bachelor of Science Electrical Engineering, Bachelor of Science Electrical Engineering at Sharif University of Technology