Curriculum Vitae

• (incoming) 98-317 Hype for Types, Spring 2025

• 15-251 Great Ideas in Theoretical Computer Science, Fall 2024
• 15-150 Principles of Functional Programming, Fall 2023
• 15-150 Principles of Functional Programming, Spring 2024

• Web Programming and Flask in 15-113 Special Topics in Applied Python Programming, Spring 2023
• Machine Learning and PyTorch in 15-113 Special Topics in Applied Python Programming, Spring 2023

• Designed and implemented an interpreter for a custom configuration language to automate data flow from API sources into a data warehouse, speeding up data ingestion workflow by 12 times and saving 10+ hours every week for the team
• Developed mechanisms for the interpreter to handle API limitations such as page size limits and recover from rate limits and network failures, allowing reliable processing of 1,000,000+ rows of data
• Provisioned cloud infrastructure using Terraform and GitHub Action to sync developer data from 2k+ repositories with one button
• Powered productivity metrics dashboards updated and viewed weekly by 20+ senior leaders, enabling data-driven decisions for 800+ developers to improve software development efficiency and code quality

• Designed and graded homework assignments on functional programming in the Standard ML programming language
• Taught labs to 30+ students and held weekly office hours to help students understand course content and appreciate functional programming

• Developed a mass spectrum database with Professor Hosein Mohimani’s team to index, cluster, and query against terabyte-scale genomics and metabolomics datasets for antibiotics discovery (Rust, 80k lines, fully documented and unit-tested code base)
• Optimized the large-scale tandem mass spectrometry data clustering algorithm Networking+ and improved accuracy of the memory-efficient mass spectrometry search tool MASST+, achieving speed ≥ 2 orders of magnitude faster than existing tools
• Implemented index data structures and caching strategies for spectrum databases while performing testing and profiling to ensure their correctness and efficiency
• Created data visualizations to analyze software performance
• One publication in Nature Biotechnology: Fast mass spectrometry search and clustering of untargeted metabolomics data

• Built blockchain testing infrastructure automation system with Electron desktop application and React front-end
• Developed internal dashboard with REST API, Redis, and JavaScript to control and monitor hundreds of nodes

• Something like making LLMs look at databases

• Pioneering novel neural techniques to improve reconstructions of protolanguage word forms, advised by Professor David R. Mortensen
• Achieved 3% and 1% accuracy improvement on Middle Chinese and Latin reconstruction respectively compared to state-of-the-art methods (Python, PyTorch, 10k lines, 4,000+ experiments on WandB)
• Developed a new neural architecture for the new NLP task semisupervised protoform reconstruction and achieved accuracy 2 to 7 percentage points higher than all baseline methods, advancing the field of computational historical linguistics
• LREC-COLING 2024 paper: Improved Neural Protoform Reconstruction via Reflex Prediction
• ACL 2024 paper: Semisupervised Neural Proto-Language Reconstruction

• Designed cellular automata based model for 3D structure generation
• Trained and tested various models on dataset (PyTorch)

• Reviewed literature on neural machine translation, information retrieval, and search ranking
• Built prototype translators and search engines

• Developed and maintained multiple online booking systems for rooms and studios on campus (JavaScript)
• Facilitated collaboration and incorporated code contributions

• Assisted participants and staff with computing-related issues
• Facilitate publicity of the program via media technology (videos, photos, etc.)

See also Portfolio

• Scrapped and cleaned webpage data to create a corpus containing information about the history, events, courses, culture, etc. of CMU, SCS, and LTI
• Developed a Retrieval Augmented Generation (RAG) system that retrieves information from the corpus to answer questions
• Annotated question-answer pairs and performed extensive experiments to test different LLMs and prompting techniques
• Achieved the best performance among all submissions in the class

• Designed the Co programming language, which allows programmers to write concurrent programs while building on top of the simplicity and safety of C0
• Implemented a compiler for Co, with support for concurrency features like coroutines, channels, and mutexes

• Designed and implemented a C0 (a memory-safe subset of C) to x86-64 compiler in Rust (in a team of 2)
• Made 400+ commits, opened 40+ pull requests, passed 10,000+ test cases

• Completed the implementation of a relational database management system (RDBMS)
• Implemented key components of the RDBMS including a buffer pool manager, a disk-backed hash index (extendible hashing), an execution engine with query optimization, and an optimistic multi-version concurrency control scheme

• Designing and creating an open-source, NLP-enhanced, content-based language learning app
• Developing algorithms and data structures for text processing (Rust), multi-lingual tokenization and lemmatization integrations (Python), back-end with REST API (Rust, Axum, SurrealDB), and front-end user interfaces (Svelte, TypeScript)

• a high-utilization dynamic memory allocator
• a shell that supports running and managing concurrent foreground and background jobs
• a multi-threaded TCP/IP proxy server with cache
• a cache simulator
• cache-optimized functions for matrix computation (performance ranked #1 in class leaderboard)

Contributed to a source-available browser-based editor, bytecode interpreter, and debugger for the C0 programming language (a safe subset of C) used by hundreds of students in CMU’s introductory C programming class. Technologies used include TypeScript and React.

Implemented and maintained online booking systems for music practice rooms and art studios on campus (JavaScript, Express.js); managed Git repository, performed code review, and incorporated code contributions.

Developed an online donation platform in collaboration with 4 student organizations in high school to assist Wuhan Union Hospital at the start of the COVID-19 pandemic (Node.js), raising a total of ¥75,000 (CNY).

• Neural Networks and Deep Learning, taught by Andrew Ng, Coursera, certificate
• Improving Deep Neural Networks: Hyperparameter Tuning, Regularization and Optimization, taught by Andrew Ng, Coursera, certificate

• Helped develop and test missing control protocol to operate the IRIS rover on the moon
• IRIS was launched to the moon in Jan. 2024 with the Peregrine Lander

• Directed IT Helpdesk to help solve tech problems on campus
• Assisted events held by other student organizations using technology
• Provided service at the IT Helpdesk for 120+ hours

• Organized, directed, and managed the three-day inter-school SCMUN X conference
• Assisted the Security-General in organizing and directing the SCMUN IX conference
• Chaired the United Nations Security Council (UNSC) at the SCMUN VIII conference
• Revised SCMUN’s design system

• Planned charity music events
• Arranged music and conducted music performances
• Entertained elderly people in the local community, helped disabled children succeed in music, and raised funds for charity organizations

• Represented students’ voices and pushed for changes to improve students’ life
• Led collaborations with other student organizations to organize events for the student body
• Designed promotional materials

* This page may be out of date