Postdoctoral Research Fellow · York University

Building trustworthy AI agents for software engineering.

I’m Jiho Shin, a postdoc at York University’s Lassonde School of Engineering. I study agentic software engineering: how humans work with agents, how agents and tools collaborate, and how reliable LLMs really are at generating, testing, and repairing code.

Scroll ↓

Research focus

From AI4SE to agentic SE

Can agents do real software engineering, can people work with them well, and can we trust what comes out?

01 · Agentic SE

Human–agent collaboration

How developers interact with coding agents, and how multiple agents and tools coordinate to finish complex engineering tasks.

02 · Trustworthy LLMs for code

Testing & evaluation

Unit test and oracle generation, retrieval-augmented testing, benchmarks for vulnerability detection, and metrics that actually mean something.

03 · AI4SE foundations

Repair, prediction, translation

Program repair, explainable defect prediction, code summarization and translation, and the quality of the ML libraries underneath.

  • 0Peer-reviewed papers
  • 0Paper awards
  • 0Preprints under review
  • 0Venues served as PC / reviewer

Selected work

Papers I’d start with

All publications ↗
  • 2026
    Retrieval-Augmented Test Generation: How Far Are We?
    Jiho Shin, Nima Shiri Harzevili, Reem Aleithan, Hadi Hemmati, Song Wang
    ICSE 2026
    ↗
  • 2026
    SecVulEval: Benchmarking LLMs for Real-World C/C++ Vulnerability Detection
    Md Basim Uddin Ahmed, Nima Shiri Harzevili, Jiho Shin, Hung Viet Pham, Song Wang
    AIware 2026🏆 Best Benchmark & Dataset Award
    ↗
  • 2026
    StaAgent: An Agentic Framework for Testing Static Analyzers
    Elijah Nnorom, Md Basim Uddin Ahmed, Jiho Shin, Hung Viet Pham, Song Wang
    TOSEM 2026to appear
    ↗
  • 2024
    Assessing Evaluation Metrics for Neural Test Oracle Generation
    Jiho Shin, Hadi Hemmati, Moshi Wei, Song Wang
    TSE 2024ICSE-JF 2025
    ↗