GPU SYSTEMS · DIGITAL DESIGN · VERIFICATION

Conrad FernandezGPU systems and
hardware acceleration.

I'm a Graphics Systems Engineer at AMD, working on GPU systems and verification. I studied computer engineering at Queen's University. Outside work I build processors and accelerators to understand how they behave.

AMDGraphics Systems Engineering
Best Overall ProjectAMD Intern Innovation Showcase · DeepDebug
2nd nationallyProgramming · Canadian Engineering Competition 2025

BUILDING NOW

Zero to NPU

Notes on building an int8 accelerator for a small language model on a Zynq FPGA: fixed-point arithmetic, a systolic MAC engine, memory movement, and sequencing.

Course notes and design targets. The interactive matrix-engine demo is available now; hardware capstones are goals, not published benchmark results.

Course notes →

SELECTED WORK

Projects

What each one does, how it is built, and what was measured.

Systolic array diagram: activations move right through stationary weights; partial sums move down.

01 / ACCELERATOR ARCHITECTURE Interactive demo

Inside a systolic array

A cycle-by-cycle view of a weight-stationary matrix engine, checked against a reference matrix multiply. Shows pipeline fill, steady state, and drain.

Seven verification configurations · int8 MACs · JavaScript + Canvas

02 / SILICON DEBUGGING Team project

DeepDebug

AI-assisted debugging for silicon validation, built with Vivek Chokkalingam during our AMD internship. It analyses system state and suggests where to look next.

Best Overall Project · AMD Intern Innovation Showcase

03 / PROCESSOR DESIGN RTL + simulation artifacts

A RISC processor in Verilog

Datapath, arithmetic, registers, memory, and I/O, with per-operation testbenches and waveform files. Coursework, built from the instruction set up.

Verilog · ModelSim · Intel Quartus · DE0-CV target

04 / EMBEDDED SYSTEMS Team prototype

LoRa wildfire sensing

A team-built ESP32 sensor network connecting radio nodes, a gateway, and a map dashboard. The documented outdoor test reached 300 metres in a semi-forested area.

ESP32 · LoRa · Supabase · Web dashboard

EXPERIENCE & RECOGNITION

Experience

Graphics Systems Engineer · AMD

GPU systems and verification. During an internship I built DeepDebug, an AI-assisted silicon debugging tool, with Vivek Chokkalingam.

LinkedIn ↗

Computer Engineering · Queen's University

Processor design, embedded systems, and team engineering projects. Our team placed second in programming at the 2025 Canadian Engineering Competition with an MRI triage prototype.

Competition result ↗

CONTACT

Get in touch

Happy to talk about GPUs, accelerators, or verification. Email is best.

Get in touch ↗