avatar
Yifeng He
We must know, we shall know.
qrcode

About Me

Hey there! I’m 贺一峰 (He, Yifeng), a third-year Ph.D. candidate at UC Davis. My advisor is Professor Hao Chen. I got my B.S. (with Honors) from UC Davis too, double majoring in Applied Math and Computer Science. During undergrad, I also served as president of HackerHub, a student-run tech club on campus.

These days, I spend most of my time thinking about how AI and software engineering can help each other out. More specifically, my research interests are:

  • analyzing and improving agent-native research from a software aspect. (arXiv 2026)
  • building and evaluating AI for software engineering, testing, and security — and using software engineering to build better AI in return. (ISSRE 2026, LMPL 2026, AST 2026, NeurIPS 2025, EMNLP 2025, ISSTA 2024, EMNLP 2024, ACL 2023, arXiv 2026 (1), arXiv 2026 (2))
  • testing things beyond code :) (ACL 2026)
  • making AI itself more secure — both foundation models and agents. (RAIE 2025, arXiv 2024)
  • safe and secure programming languages (bonus points if useful for AI)
    • PS: useful PL for AI is a very unclear topic, and should be researched on. I need to work on this LOL.
  • AI for art, especially music.
    • I am mostly interested in using AI to understand the commons and differences between genres, and generate music with genre transfer.
    • See my early work here.
    • Please reach out if you are interested in collaborating on this topic!

I’m also into video games. Some of my all-time favorites include Pokémon (recently played Sword 🙃, Violet, and Arceus 😊), The Witcher 3, Clash of Clans (hit world rank #169 back in April 2018!), and Genshin Impact (though I haven’t touched it since 4.4, thanks miHoYo 😅). Other games I’ve enjoyed or recommend: The Legend of Zelda series, Palworld (was obsessed for a while… then drifted back to egg hatching in Pokémon).

I also enjoy powerlifting. I have trained for a little over two years. My (training) personal records are Squat 405lb (184kg), Bench Press 245lb (111kg), Deadlift 455lb (206kg) summing up to 1105lb (501kg) @ 85kg body weight. I also dabble in photography. If you’re curious, feel free to check out some of my shots on Instagram.

Curriculum vitae [PDF]

Education

Ph.D. in Computer Science, UC Davis (2023 – present)

B.S. in Computer Science and Applied Mathematics with Honors, UC Davis (2019 – 2023)

Publications

(* equal contribution)

2026

Yifeng He, Jicheng Wang, Yinzhe Zhao, Chengyang Shi, Jiachen Liu, Hao Chen. The Greatness of Science Cannot Be Planned: Agentic Auto-Research is Fuzz Testing. [arXiv], [blog], [code], [trajectories].

Ruogu Yang*, Yifeng He*, Yundi Xu, Yuqing Wei, Hao Chen. LLM-Based Invariant Testing for Software Functional Bugs. International Symposium on Software Reliability Engineering (ISSRE), 2026. [DOI], [arXiv], [PDF], [code].

Yifeng He, Yinzhe Zhao, Jicheng Wang, Hao Chen. Is Progressive Disclosure All You Need for Long-Context Agents?. [arXiv].

Jicheng Wang, Yifeng He, Zili Wang, Hanwen Xing, Arkaprava De, Hao Chen. ContractBench: Can LLM Agents Preserve Observation Contracts?. [arXiv], [code].

Yifeng He, Yundi Xu, Christopher Castro Gaw Gonzalo, Zili Wang, Hao Chen. Invariant Pretraining for Robust Code Representations. International Workshop on Language Models and Programming Languages (LMPL), 2026. [DOI], [PDF], [arXiv], [code],

Yifeng He, Ziye Tang, Hao Chen. Content Fuzzing for Escaping Information Cocoons on Digital Social Media. Findings of Association for Computational Linguistics (ACL), 2026. [DOI], [PDF], [code], [slides], [poster].

Xiangyi Li, Wenbo Chen, Yimin Liu, Shenghan Zheng, Xiaokun Chen, Yifeng He, …+34 authors. SkillsBench: Benchmarking How Well Agent Skills Work Across Diverse Tasks, 2026. [arXiv], [website], [code].

Hongxiang Zhang, Yuyang Rong, Yifeng He, Hao Chen. LLAMAFUZZ: Large Language Model Enhanced Greybox Fuzzing. International Conference on Automation of Software Test (AST), 2026. [arXiv], [code].

2025

Yifeng He, Luning Yang, Christopher Castro Gaw Gonzalo, Hao Chen. Evaluating Program Semantics Reasoning with Type Inference in System F. Neural Information Processing Systems Datasets & Benchmarks Track (NeurIPS), 2025. [OpenReview], [PDF (recommended)], [arXiv], [code], [slides], [poster],

Yifeng He, Jicheng Wang, Yuyang Rong, Hao Chen. FuzzAug: Data Augmentation by Coverage-guided Fuzzing for Neural Test Generation. Conference on Empirical Methods in Natural Language Processing: Findings (EMNLP), 2025. [DOI], [PDF], [arXiv], [code], [video (YouTube)], [video (bilibili)] [slides], [poster].

Yifeng He, Ethan Wang, Yuyang Rong, Zifei Cheng, Hao Chen. Security of AI Agents, International Workshop on Responsible AI Engineering (ICSE-RAIE), 2025. [DOI], [PDF], [arXiv (recommended)], [code], [slides].

2024

Yifeng He, Jiabo Huang, Yuyang Rong, Yiwen Guo, Ethan Wang, Hao Chen. UniTSyn: A Large-Scale Dataset Capable of Enhancing the Prowess of Large Language Models for Program Testing, International Symposium on Software Testing and Analysis (ISSTA), 2024. [DOI], [PDF], [code], [slides], [poster].

Jiabo Huang, Jianyu Zhao, Yuyang Rong, Yiwen Guo, Yifeng He, Hao Chen. Code Representation Pre-training with Complements from Program Executions. Conference on Empirical Methods in Natural Language Processing: Industry Track (EMNLP), 2024. [DOI], [PDF], [poster].

2023

Jianyu Zhao, Yuyang Rong, Yiwen Guo, Yifeng He, Hao Chen. Understanding Programs by Exploiting (Fuzzing) Test Cases, Findings of Association for Computational Linguistics (ACL), 2023. [DOI], [PDF], [code].

Others :)

Yifeng He, Ethan Wang, Jicheng Wang, Xuanxin Ouyang, Hao Chen. Code Generation by Differential Test Time Scaling. [arXiv], [code].

Hongxiang Zhang, Yifeng He, Hao Chen. SteerDiff: Steering towards Safe Text-to-Image Diffusion Models. [arXiv], [code].

Jicheng Wang, Yifeng He, Hao Chen. RepoGenReflex: Enhancing Repository-Level Code Completion with Verbal Reinforcement and Retrieval-Augmented Generation. [arXiv].

Yifeng He, Big Data and Deep Learning Techniques Applied in Intelligent Recommender Systems, International Conference on Civil Aviation Safety and Information Technology (ICCASIT), 2022. [DOI].

Services

Organizing Committee / Co-Chair for:

Program Committee Member / Reviewer for:

  • ACM Conference on AI and Agentic Systems (CAIS), Main & Demo: 2026
  • Conference on Language Modeling (COLM): 2026
  • International Conference on Learning Representations (ICLR): 2026, 2025
  • Neural Information Processing Systems (NeurIPS): 2026
  • International Conference on Artificial Intelligence, Computer, Data Sciences and Applications (ACDSA): 2026
  • International Conference on Multimedia (MM): 2026, 2025
  • International Conference on Software Engineering Advances (ICSEA): 2025
  • Association for Computational Linguistics LLM Security Workshop (ACL-LLMSEC): 2025

Artifact Evaluation Committee for:

  • IEEE Symposium on Security and Privacy (S&P): 2026
  • Conference of the ACM Special Interest Group on Data Communication (SIGCOMM): 2025

Honors and Awards

Citation For Outstanding Performance, Department of Mathematics, UC Davis

Dean’s Honor List, Fall 2019, Spring 2020, Spring 2021, Spring 2022, College of Letter & Science, UC Davis

Current Side Projects

SkillsHub – Homebrew for AI Agent Skills. Install, manage, and share skills across AI coding agents with a unified CLI. [website], [code].

Teaching

  • ECS 36C Data Structure
    • Teaching Assistant. Fall 2023, Spring 2023, Fall 2024.

Industry Experience

Aamzon Web Services, Kiro Science. Applied Scientist Intern. 06/2026 - 09/2026

  • Make agents smarter on analyzing code.

ByteDance (Toutiao), Creator Income Platform. Software Development Engineer Intern. 04/2021 - 07/2021

  • Used microservice tech to connect parts of the author income settlement bushiness
  • Transformed author-relation data architecture design from relational database (SQL) to graph database (Gremlin) to
  • low better efficiency for the business model
  • Refactored the income calculation control process with in Python 3

ByteDance (Xigua Video), Creator Experience Team. Software Development Engineer Intern. 07/2021 - 08/2021

  • Created a data cleaner with ORM to maintain the size and readability of online data settlement table
  • Created the offline flow of Medium-Length Video Encouragement Project for weekly data calculation
  • Built the interface for frontend of web and mobile app to display the data visualization
Past Toy Projects

CourseReco 06/2022 - 09/2022

  • Designed the overall system architecture
  • Led the programming for API server and recommender engine
  • Negotiated with the third-party provider, SchedGo, for data service
  • Provided technical leadership to teammates

Music Genre Classifier 05/2022 - 06/2022

  • Processed music samples into spectrogram by Short-time Fourier transform
  • Designed the appropriate model (CNN) to classify spectrograms into category
  • Analyzed the resulting model and test outputs with saliency maps

ImageOrientation 03/2022 - 04/2022

  • Pre-processed image data by rotating them with random generated angles, and assigned these angles as label
  • Designed the appropriate CNN for regression task, tested and improved the model
  • Applied Hyper-parameter tuning based on train, validation, and tested results

Dcash-server 05/2021 - 07/2021

  • Created a multi-threaded API server using C++ to allow users to create accounts to make deposit and transfer
  • Used MySQL to store and maintain user data
  • Made API calls to the Stripe API server to handle credit card information

Genshine Impact Gacha Analyzer 08/2021 - 09/2021

  • Designed fetching process of gacha data from MiHoYo and categorized the process
  • Stored data into local database automatically, wrote into excel for data analysis by option
  • Generated text or graph visualization report from data analyze results
Activities

HackerHub Club, UC Davis 07/2020 - 06/2023

Co-founder, President, Technical Officer

  • Design and maintain a course recommendation system, CourseReco, for UC Davis students
  • Organize and lead the Code Jam Competition on data visualization, AI, augmented reality and virtual reality, and machine learning
  • Coach in introductory programming workshops, including Assembly, functional programming, recommender system, generative adversarial network, etc.