About me

I am currently a Ph.D. student candidate (Dec, 2023) at ECE@Purdue. I am supervised by Dr. James Davis.

My research interest is focused on SE4AI, and Responsible AI. I currently work on advancing AI safety/security and utility, more specifically on improving reusability, trustworthiness, and security of pre-trained deep learning models (PTMs) from model registries/hubs, such as Hugging Face.

Here is a list of projects that I am actively working on:

  • Trustworthiness: Anomaly detection for PTM naming and architectural defects.
  • Reusability: PTM recommendation system considering engineering requirements.
  • Security: Defense against Pickle deserialization vulnerabilities in PTMs.

My work has been published at MSR’24, ICSE’23, MSR’23, ESEC/FSE’22, ASE’22, and SCORED’22.

I am also the leader of Purdue’s Pre-Trained Models in Deep Learning VIP team (previously TensorFlow Model Garden VIP team where we collaborated with Google).