Jiacheng Liang

About Me

I am a Ph.D. candidate in Computer Science at Stony Brook University, advised byProf. Ting Wang. My research focuses on LLM safety and trustworthy AI, with an emphasis on identifying security vulnerabilities and developing robust defenses.

My work has been published at top-tier venues including IEEE S&P, ACM CCS, ACL, EMNLP, ICML, ICLR, and AsiaCCS. I am currently an Applied Scientist Intern at Prime Video & Amazon MGM Studios, working on reinforcement learning for LLM Recommendation. Previously, I interned at Amazon AGI (Nova Responsible AI), where I developed adaptive red-teaming methods for RLHF systems.

My research interests include LLM post-training and safety alignment, LLM agent safety, LLM deployment security, jailbreak and backdoor defense, and RAG safety.

Selected Publications (Full List)

No publications match this combination.

• ARES: Adaptive Red-Teaming and End-to-End Repair of Policy-Reward System Jiacheng Liang, Yao Ma, Tharindu Kumarage, Satyapriya Krishna, Rahul Gupta, Kai-Wei Chang, Aram Galstyan, Charith Peris Annual Meeting of the Association for Computational Linguistics (ACL 2026) Alignment & Post-TrainingAttacks & Red-Teaming
• AutoRAN: Weak-to-Strong Jailbreaking of Large Reasoning Models Jiacheng Liang, Tanqiu Jiang, Yuhui Wang, Rongyi Zhu, Fenglong Ma, Ting Wang Annual Meeting of the Association for Computational Linguistics (ACL 2026) Reasoning ModelsAttacks & Red-Teaming
• GraphRAG under Fire Jiacheng Liang, Yuhui Wang, Changjiang Li, Rongyi Zhu, Tanqiu Jiang, Neil Gong, Ting Wang IEEE Symposium on Security and Privacy 2026 (IEEE S&P 2026) RAGAttacks & Red-Teaming
• AgentLAB: Benchmarking LLM Agents against Long-Horizon Attacks Tanqiu Jiang, Yuhui Wang, Jiacheng Liang, Ting Wang International Conference on Machine Learning (ICML 2026) LLM AgentsAttacks & Red-TeamingBenchmarking
• MAGE: Safeguarding LLM Agents against Long-Horizon Threats via Shadow Memory Yuhui Wang, Tanqiu Jiang, Jiacheng Liang, Charles Fleming, Ting Wang ACM Conference on Computer and Communications Security (ACM CCS 2026) LLM AgentsSafety Defense
• RASA: Routing-Aware Safety Alignment for Mixture-of-Experts Models Jiacheng Liang, Yuhui Wang, Tanqiu Jiang, Ting Wang Empirical Methods in Natural Language Processing (EMNLP 2026) Alignment & Post-TrainingSafety Defense
• WaterPark: A Robustness Assessment of Language Model Watermarking Jiacheng Liang, Zian Wang, Lauren Hong, Shouling Ji, Ting Wang Empirical Methods in Natural Language Processing (EMNLP 2025) WatermarkingAttacks & Red-TeamingBenchmarking
• Model Extraction Attacks Revisited Jiacheng Liang, Ren Pang, Changjiang Li, Ting Wang Asia Conference on Computer and Communications Security (Asia CCS 2024) Attacks & Red-TeamingBenchmarking
• Data to Defense: The Role of Curation in Customizing LLMs Against Jailbreaking Attacks Xiaoqun Liu*, Jiacheng Liang*, Luoxi Tang, Muchao Ye, Weicheng Ma, Zhaohan Xi Empirical Methods in Natural Language Processing (EMNLP 2025) Alignment & Post-TrainingSafety Defense
• LaCache: Robust Semantic Caching for LLM Serving Jiacheng Liang, Yuhui Wang, Tanqiu Jiang, Ting Wang LLM ServingSafety Defense

Services

Program Committee/Reviewer

International Conference on Machine Learning (ICML, 2026 Gold)
Neural Information Processing Systems (NeurIPS)
International Conference on Learning Representations (ICLR)
Association for the Advancement of Artificial Intelligence (AAAI)
Transactions on Machine Learning Research (TMLR)
IEEE Transactions on Dependable and Secure Computing (TDSC)
IEEE Transactions on Information Forensics and Security (TIFS)
IEEE Internet of Things Journal
Transactions on Intelligent Systems and Technology (TIST)
Cybersecurity Springer

Internships

Powered by Jekyll and Minimal Light theme.