About

Test vs Mutant: Adversarial LLM Agents for Robust Unit Test Generation

Pengyu Chang, Yixiong Fang, Silin Chen, Yuling Shi, Beijun Shen, and Xiaodong Gu.

ISSTA 2026 · arXiv

Audio Language Model for Deepfake Detection Grounded in Acoustic Chain-of-Thought

Runkun Chen, Yixiong Fang, Pengyu Chang, Yuante Li, Massa Baali, and Bhiksha Raj.

arXiv preprint · arXiv

LMTransplant

Contributed to task selection, synthetic-data prompt design, and evaluation for this text data augmentation project, accepted at EMNLP 2025.

Google

Software Engineering Intern · San Francisco

Built a synthetic-data pipeline for Gemini content safety using structured policy taxonomies and conditional generation, then evaluated downstream classifiers.

Electronic Arts

Software Engineer Intern · Shanghai

Integrated Kiali into a Grafana and Prometheus monitoring stack and reduced Kiali data volume by 85%, resolving memory bottlenecks.

Shanghai Jiao Tong University

Research Assistant · Shanghai

Worked on few-shot NLP data augmentation and multi-agent LLM methods for unit-test generation.

Minik8s

A small container orchestration system in Go with control loops, horizontal autoscaling, API-object management, and NFS-backed persistent storage.

LSM-KV

A C++ key-value store using skip lists, Bloom filters, and an LSM-tree design.

Code

Carnegie Mellon University

M.S. in Information Networking · GPA 4.0/4.0

Teaching Assistant, Introduction to Deep Learning.

Shanghai Jiao Tong University

B.E. in Software Engineering.