PhD candidate working on self-aware AI models.
HPI · Stanford · San Francisco Bay Area
About
I am a PhD researcher in the Artificial Intelligence group at the Hasso Plattner Institute (HPI), advised by Prof. Gerard de Melo, and currently a visiting researcher at Stanford University. My research centers on epistemic uncertainty and metacognition in large language models - building methods that let models recognize, quantify, and express the limits of their own knowledge, and studying how they reason about what they know. This spans factuality and hallucination reduction, calibrated abstention, knowledge editing, interpretability, and reasoning, across pretraining, alignment, and inference.
I am currently working on a few new projects and am open to collaborations. If your interests overlap with mine, please feel free to get in touch.
Research
- Epistemic uncertainty and metacognition in LLMs
- Factuality and hallucination
- Reasoning and interpretability
- Knowledge in language models, knowledge editing
- Alignment and instruction tuning
- LLM safety
News
- Aug 2026Began a visiting research position at Stanford University.
- 2025InFact accepted to Findings of EMNLP 2025.
- 2025Featured in The Wall Street Journal on why AI chatbots struggle to admit “I don’t know.”
- 2024I Don’t Know: Explicit Modeling of Uncertainty with an [IDK] Token accepted to NeurIPS 2024.
- 2024Evaluating the Ripple Effects of Knowledge Editing in Language Models published in TACL.
Selected work
- InFact: Informativeness Alignment for Improved LLM Factuality Findings of EMNLP 2025
- I Don't Know: Explicit Modeling of Uncertainty with an [IDK] Token NeurIPS 2024
- Evaluating the Ripple Effects of Knowledge Editing in Language Models TACL 2024
- LM vs LM: Detecting Factual Errors via Cross-Examination EMNLP 2023
- Crawling the Internal Knowledge-Base of Language Models Findings of EACL 2023