Mark Sui - RTL Design Engineer Portfolio

ASIC / FPGA / Software Engineering

UCSD · Math · CE · ECE.

UCSD
MATH
CE
ECE

I'm Mark Sui, an ECE master's student at UC San Diego with a mathematics and computer engineering foundation. I work from cycle-level RTL and FPGA implementation to software tools that make complex systems easier to build and verify.

Seeking ASIC / FPGA / Software Engineering roles.

View RTL Projects Resume Technical Work GitHub Get in Touch

Target Role ASIC / FPGA / SWE

HDL SystemVerilog / Verilog

Design Flow Simulation / FPGA

Implementation Hardware / Software

Open to ASIC, FPGA, RTL design, and software engineering opportunities. Start with the Viterbi decoder for hardware depth, then review the public tools for software and systems range.

View RTL Projects Resume GitHub Contact

RTL Signal Flow

01 / SPECIFICATION Start with behavior that can be stated cycle by cycle. Interfaces · latency · constraints

Selected RTL Projects

Direct RTL Evidence Digital Design / SystemVerilog / FPGA

FPGA-Based Viterbi Decoder

Designed a convolutional encoder and modular Viterbi decoder with branch-metric, add-compare-select, and traceback units aligned cycle by cycle.

Architecture: Decomposed the decoder into BMC, ACS, and Traceback modules.
Built With: SystemVerilog, FPGA-oriented RTL, simulation, waveform debugging.
Verification: Used controlled error cases to keep path metrics and traceback behavior traceable.

Project summary Resume context

Logic and CMOS Studio interface showing the Boolean formula workspace and gate-level circuit diagram

Public Design Tool Logic + CMOS / Verilog / React

Logic & CMOS Studio

Browser-based tool connecting Boolean expressions to truth tables, K-maps, simplified logic, generated Verilog, and static CMOS schematics.

Role: Designed and built the end-to-end logic exploration workflow.
Built With: React, TypeScript, Boolean logic, Verilog, static CMOS.
Why It Matters: Makes the transformation from logic specification to implementation inspectable.

Live app Source

Report-Backed RTL Research VerilogEval / Python / RTL Debugging

MAGE RTL Generation

Extended a multi-agent Verilog generation flow with failure routing, route-specific repair prompts, edit guards, and task-local debugging memory.

Problem: Generated RTL can compile yet still fail from interface, timing, or behavioral mismatches.
Built With: Python, VerilogEval-V2, Icarus Verilog, routed repair workflows.
Evidence: Public source and a report documenting the routing and memory experiments.

Report PDF Source

Experience

Jul — Aug 2026 Data Quality Reviewer AfterQuery Jun 2025 — Jun 2026 Research Fellow Handshake Jun — Dec 2024 Embedded Systems Intern Shandong Haike

View complete experience →