PhD student in Computer Science at The Hebrew University of Jerusalem, working on how we evaluate large language models, and how to make them trustworthy for empirical legal research.
I am a PhD student in Computer Science at the Hebrew University of Jerusalem, advised by Prof. Gabriel Stanovsky and Prof. Renana Keydar. My research focuses on Natural Language Processing, developing evaluation methodologies for LLMs that address challenges like prompt sensitivity and benchmark reliability, alongside work on extracting structured information from large document collections, with applications in the legal domain (yes, both my parents are jurists, so my friends tell me these are daddy-mommy issues).
I received my M.Sc. in Computer Science (2021–2023) under the supervision of Prof. Gabriel Stanovsky, where my thesis research applied computational methods to analyze judicial attitudes in court decisions about sexual violence cases. I also hold a B.Sc. in Mathematics and Computer Science (2017–2019, graduated with honours) from the Hebrew University of Jerusalem.
Beyond academia, I've gained industry experience as an Algorithm Developer at OrCam for two years, a Data Scientist Intern at Microsoft, and a summer research internship at Lightricks.
I'm always open to collaborations and enjoy discussing research at the intersection of NLP, legal AI, and evaluation methodologies. Feel free to reach out! ✉️
Princeton Language and Intelligence (PLI) Lunch Series, Princeton, NJ
October 15, 2026
A talk presenting ScheMatiQ, a system that turns a research question into a structured dataset by discovering a schema over a document corpus.
AI Measurement Science Workshop (AIMS) at COLM 2026, San Francisco, CA
October 9, 2026
A talk presenting our COLM 2026 paper on extending LLM benchmarks with new items and models without re-running full evaluations.
TopResearch by MDLI, Tel Aviv, Israel
September 14, 2026
A talk presenting our ACL 2025 paper on LLM prompt sensitivity, and the follow-up work on PromptSuite and ReliableEval.
ISCOL 2025, the Israeli Seminar on Computational Linguistics, Bar-Ilan University, Israel
December 18, 2025
An oral presentation of our ACL 2025 paper on LLM prompt sensitivity.
Graduate course, The Hebrew University of Jerusalem
2025–2027
A graduate course for about 60 students, taught by Prof. Roy Schwartz and Prof. Gabriel Stanovsky, covering cutting-edge NLP research such as efficiency, multilingual NLP, and evaluation. Students carry out semester-long research projects of their choice, guided by the course staff from the initial idea to the final report.
Developed during my time at Microsoft: