☕️Research Engineer
About Me
Hello:) I’m a research engineer in ELLIS Institute Tübingen with the OpenEuroLLM project. I’m broadly interested in ML for social good. This includes measuring emergent behaviours / their impact on society, monitoring its trustworthiness and safety, and aligning it with human values.
My work sits at the intersection of model behavior and safety:
- Interaction & Behavior: Currently interning with Jonas Geiping, where I focus on understanding how models interact with their environments and users.
- Safety & Robustness: Previously collaborated with Professors Mengyue Yang, Linyi Yang and Jun Wang on web agent safety and mitigating spurious correlations in LLMs.
- Optimization: For my thesis, I had the privilege of working with Pu-Jen Cheng and Jyun-Yu Jiang on Augmenting Retrieval-Augmented Generation (RAG) with Passage Combination.
When I’m free, I enjoy skateboarding, cooking, hopping into local cafes, and exploring the city 🌞💫
- Machine Learning
- Alignment
- Trustworthy AI
- AI safety
- Causal Reasoning
- Information Retrieval
MSc Computational Statistics and Machine Learning
University College London
MSc Computer Science
National Taiwan University
BSc Electrical Engineering and Computer Science, EECS Honors Program
National Chiao Tung University (Now National Yang Ming Chiao Tung University)
My research interests lie in trustworthy & interpretable AI, especially in understanding LLM reasoning by causality. Before joining UCL, I worked in research intern roles at MediaTek Research UK and Yahoo!, where I gained experience in generative AI, machine learning (ML), natural language processing (NLP), retrieval-augmented generation (RAG), and model understanding.
Please reach out to collaborate 😃