Josh Ludan
PhD student · Computer and Information Science · University of Pennsylvania
I work on natural language processing, especially on making large language models more robust and interpretable. I’m advised by Mark Yatskar and Chris Callison-Burch, and I’m a 2026 ASSET Center AWS Fellow.
News
- Apr 2026
Named a 2026 ASSET Center AWS Fellow at Penn Engineering.
- 2025
Medex published at NeurIPS 2025.
- Jan 2025
Started my PhD in Computer and Information Science at Penn.
- 2024
Best Thesis Award, Penn MSE in Data Science, for my thesis on Text Bottleneck Models, interpretable-by-design text classifiers.
- 2023
Best Practicum Award, Penn MSE in Data Science.
Publications
- 2025
- 2024
- 2023
- 2023
Experience
- 2025 – now
PhD Student, Computer and Information Science · University of Pennsylvania
Started in January 2025, advised by Mark Yatskar and Chris Callison-Burch; 2026 ASSET Center AWS Fellow. Before the PhD, I worked in the same lab from 2022 during my combined undergraduate and accelerated master’s program, on robust, interpretable LLM systems (Best Practicum Award 2023; Best Thesis Award 2024 for my thesis on Text Bottleneck Models).
- 2022
Data Science Intern · Farallon Capital Management
Used LLMs to gauge the financial sentiment of online communities; the model outperformed every publicly available financial-sentiment model during its evaluation period.
- 2021 – 2024
VP of Projects · Penn Data Science Group
Led consulting projects for Flagler Health and EmployAI and coordinated the club’s Kaggle team.
- 2021 – 2022
Research Assistant · Penn Computational Social Science Lab
Topic models of American YouTube consumption, with Homa Hosseinmardi and Duncan Watts.
Education
- 2025 –
PhD, Computer and Information Science · University of Pennsylvania
- 2024
MSE, Data Science · University of Pennsylvania
- 2024
BSE, Networked and Social Systems Engineering · University of Pennsylvania