Ph.D. Research Proposal: Yiting Wang

Tuesday, November 25, 2025
10:00 a.m.
AVW 2460
Sarah Pham
301 473 2449
spham124@umd.edu

ANNOUNCEMENT: Ph.D. Research Proposal Exam

 

Name: Yiting Wang

 

Committee:

Professor NAME (Chair) : Gang Qu

Professor NAME: Ang Li

Professor NAME: Cunxi Yu

 

Date/time: November 25, 2025 from 10 am -1 pm

 

Location: AVW 2460

 

Title: Toward Intelligent Automation of LLM-Based Hardware Design

 

Abstract: This proposal presents a comprehensive Large Language Model (LLM)-based hardware design flow by integrating generation, optimization, and verification. Traditional hardware design requires engineers to manually write RTL code, tune optimization parameters from synthesis feedback, craft Tcl scripts, write verification scripts, and coordinate multiple Electronic Design Automation (EDA) tools —an effort that is highly time-consuming and error-prone. Our research shows how LLMs can serve as the central intelligence orchestrating the design process, producing satisfying results with minimal human intervention. The proposed flow integrates four LLM-powered components: VeriReason – a reinforcement learning–enhanced RTL generation system with reasoning capabilities, achieving 83.1% functional correctness through testbench feedback. SymRTLO – a LLM-driven optimization framework achieving up to 43.9% power, 62.5% timing, and 51.1% area improvements via symbolic reasoning. HADA – an automated hardware security verification system using multi-source data augmentation. MCP4EDA – a Model Context Protocol (MCP)–based backend-aware synthesis TCL script optimizer achieving 15–30% timing and 10–20% area improvements through closed-loop feedback. Experimental results show that this integrated approach not only surpasses fixed EDA scripts in design quality but also drastically reduces development time and manual effort. This work establishes a new paradigm in hardware design automation, where intelligent LLM agents collaborate with EDA tools to deliver a seamless, efficient, and democratized path to high-quality hardware.

Audience: Graduate  Faculty 

remind we with google calendar

 

November 2025

SU MO TU WE TH FR SA
26 27 28 29 30 31 1
2 3 4 5 6 7 8
9 10 11 12 13 14 15
16 17 18 19 20 21 22
23 24 25 26 27 28 29
30 1 2 3 4 5 6
Submit an Event