Portrait
Yiran Hu
PhD Student in Purdue CS · AI for Software Engineering · LLM Agents

About

I am a third-year CS PhD student at Purdue University advised by Prof. Lin Tan. My research focuses on applying Artificial Intelligence techniques to Software Engineering. Specifically, I develop AI-driven methods, such as LLM Agents, for code generation and other SE tasks, aimed at accelerating implementation and improving reliability.

    AI Software Engineering Large Language Models LLM Agents
Highlights
TENET: Leveraging Tests Beyond Validation for Code Generation
Pass@1: 69.08% (REPOCOD), 81.77% (RepoEval), outperforming best agentic baselines by 9.49 and 2.17 pp
REPOCOD: A Code Generation Benchmark with Repo-level Context
980 func generation tasks · 11 real-world projects · Longest canonical solution (331.6 tokens) · highest cyclomatic complexity (9.00)
SELP: Generating Safe and Efficient Task Plans for Robot Agents with LLMs
ICRA 2025 Best Paper Award Finalist · Equivalence Voting · Constrained Decoding · Domain-Specific Fine-Tuning

Publications

TENET: Leveraging Tests Beyond Validation for Code GenerationPreprint under review for publication
Yiran Hu, Nan Jiang, Shanchao Liang, Yi Wu, Lin Tan

A test-driven LLM agent for repository-level code generation that combines a Test Harness Mechanism (THM), a tailored toolset for retrieval & debugging, and a reflection-based refinement workflow (RRW). Pass@1: 69.08% (REPOCOD), 81.77% (RepoEval), outperforming best agentic baselines by 9.49 and 2.17 pp.

Test-Driven DevelopmentLLMsAgentCode Generation
Can Language Models Replace Programmers for Coding? 🐟 RepoCod Says 'Not Yet'ACL 2025
Shanchao Liang, Nan Jiang, Yiran Hu, Lin Tan

REPOCOD, a code generation benchmark with 980 problems collected from 11 popular real-world projects. In our evaluations of ten LLMs, none of them achieve more than 30 pass@1 on REPOCOD, indicating the necessity of stronger LLMs to assist developers in real-world software development.

LLMsCode Generation benchmark
SELP: Generating Safe and Efficient Task Plans for Robot Agents with Large Language Models ICRA 2025 Best Paper Award Finalist
Yi Wu, Zikang Xiong, Yiran Hu, Shreyash S. Iyengar, Nan Jiang, Aniket Bera, Lin Tan, Suresh Jagannathan

SELP (Safe Efficient LLM Planner), a novel approach for generating plans that adhere to user-specified constraints while optimizing for time-efficient execution. In drone navigation, SELP outperforms SOTA LLM planners by 10.8% in safety rate and by 19.8% in plan efficiency. For robot manipulation, SELP achieves a 20.4% improvement in safety rate.

RoboticsArtificial IntelligenceFormal Languages and Automata Theory
LocRDF: An Ontology-Aware Key-Value Store for Massive RDF Data WISA 2022
Yunxiao Zhang, Jinghan Li, Xueyang Liu, Rong Cheng, Nan Jiang, Yiran Hu, Xin Wang

LocRDF, a novel storage system for RDF graphs via key-value store supporting ontology reasoning. LocRDF achieves high storage efficiency (32.9% of Jena, 19.8% of AllegroGraph on LUBM30) and cuts query time by up to 98% over Jena and ~50% over AllegroGraph on complex queries.

RoboticsArtificial IntelligenceFormal Languages and Automata Theory

Mentoring

CS176 Data Engineering In Python
Teaching Assistant · Purdue University · 2023 fall, 2024 spring, 2025 fall
  • Office hours, grading, lab/lecture instructions
CS253 Data Structures And Algorithms For DS/AI
Teaching Assistant · Purdue University · 2025 spring
  • Office hours, grading, PSO instructions
CS251 Data Structures
Teaching Assistant · Purdue University · 2024 fall
  • Office hours, grading

Education

  • Ph.D in Computer SciencePurdue University2023.08 - Now
  • B.S. in Computer Science and TechnologyTianjin University2019.09 - 2023.05

Honors

  • Best Paper Award FinalistICRA2025
  • Outstanding GraduateTianjin University2023
  • First Prize of Tianjin UniversityChallenge Cup National College Students Science and Technology Competition2022
  • First Class ScholarshipTianjin University2021
  • Advanced Individual for Academic PerformanceTianjin University2021
  • Great TeamSummer Social Activity, Tianjin University2020
  • Special PrizeFLTRP·ETIC Cup2019

Personal

Beyond research, I'm an easy-going person who enjoys hiking, playing basketball, and capturing moments with my camera. Feel free to connect with me on Instagram if you'd like to know me outside academia!