Devin Yeung
github.com/Devin-Yeung
devin.cg.yeung.dev [at] outlook.com
blog
Education
BUN - HKBU United International College 2021.09 - 2025.06 (expected)
Bachelor’s Degree in Computer Science and Technology
• cGPA 3.93 / 4.00. Major cGPA 3.98 / 4.00. Rank (1 / 147).
• Dean’s Award in 2023-2024 (Top 0.375% schoolwide) • Dean’s Award in 2021-2022 (Top 0.375% schoolwide)
• National Scholarhsip in 2023-2024 (0.04% nationwide) • First Class Scholarship in 2022-2023
Open Source Contribution
OXC Javascript Compiler Byte Dance Web-Dev Team
• Contributed more than 3k lines of code to the linter component of the OXC named oxlint , involving:
‣ Implemented ESLint Rules: e.g. no-thenable , no-unnecessary-await , ban-ts-comment
‣ Improved the Continuous Integration: include bug fixes and new features to enhance the development experience.
Bnfgen Slides Undergraduate Graduation Project
• A general purposed BNF grammar generator, which can generate random strings from a given BNF grammar.
• Enhanced the standard BNF grammar with custom extensions, featuring:
Weighted productions for probabilistic generation
Seamless integration of regular language
Progressive type system for production rules
Variable tracking with declaration and reference
•
Rich semantic analysis: include duplicated rules, unreachable rules and dead loop rules detection.
Loxide: A Lox Interpreter in Rust Personal Project
• Developed a Lox interpreter from scratch, including a lexer, recursive descent parser, evaluator, and type checker.
• Adopted Rust’s exemplary error reporting techniques to generate clear, human-readable error messages.
Context-free Tools Personal Project
• Developed a comprehensive suite of tools for analyzing context-free grammars, including common algorithms like
First, Follow, LR(0) set construction and SLR parsing table construction.
• Focused on educational purposes, the toolkit includes detailed visualizations and step-by-step explanations.
Competition
Wordle killer: The counteraack of information theory 2023.02
• Analyzed the daily results of the Wordle puzzle from the New York Times ( thesis ).
‣ Used FB-Prophet model and Monte Carlo Simulation to predict the future distributions of Wordle result.
‣ Developed an algorithm based on information entropy to measure the diiculty level of a Wordle puzzle.
•
Awarded the Finalist (Top 1.8% worldwide) in 2023 Mathematical Contest in Modeling (aka MCM / ICM).
Internship
Teaching Assistant of Compiler Construction Course (2024 Fall) 2024.09 - Now
• Designed the BNF grammar and implemented the autograder (automatic grading system) for the course project.
• Developed tools to generate parser test cases and a dedicated playground for students to beer understand the project.
• Maintained a high-performance code similarity checker to catch students’ plagiarism behavior. (open source soon)
Project
Modeling and Control of 2-DOF Robot Arm Course Project
• Investigated control aspects of a two-degree-of-freedom robotic arm, emphasizing PID control, kinematics, dynamics,
and controller design. Developed a MATLAB simulation to verify the correctness of algorithms. ( thesis )
Skills
• Technical Skills:
‣ Programming Skills: Not limited to specific language, proficient in Rust, familiar with Python, OCaml and C.
‣ Tools: Comfortable with linux environment, familiar with git, skilled in leveraging AI.
•
Domain Specific Skills
‣ Parser: Comfortable with parser combinators and parser generators, familiar with creating hand-wrien parsers.
‣ Formal Languages: Proficient in regular expressions, context-free grammars, and automata theory.
