Scholar
Fabio J. Fehr
Google Scholar ID: WaZWY0wAAAAJ
PhD candidate, EPFL, Idiap Research Institute
Citations & Impact
Citations
41
H-index
2
i10-index
2
Publications
7
Co-authors
18
list available
Resume
Academic Achievements
- August 2025: Participated in Idiap Create Challenge, creating an AI model for musicians. July 2025: Successfully defended PhD at EPFL. June 2025: Submitted doctoral thesis 'Nonparametric Variational Information Bottleneck: Attention-based Architectures as Latent Variable Models'. May 2025: Published paper 'CoRet: Improved Retriever for Code Editing', Conference: ACL 2025.
Research Experience
- Worked at Amazon Berlin on retrieval-augmented generation (RAG) and large language models (LLMs) for code generation, Collaborators: Prabhu Teja and Giovanni Zappella.
Education
- PhD: EPFL & Idiap Research Institute, Supervisor: James Henderson, Research Area: Connection between deep attention-based models and Bayesian nonparametrics for NLP. Bachelor's and Master's: University of Cape Town, Specialization: Mathematics, Computer Science, and Statistics.
Background
- Research Interests: Artificial Intelligence, Natural Language Processing. Professional Field: Mathematics, Computer Science, and Statistics. Summary: Believes in leading by example, values honesty, humility, and kindness. Passionate and tenacious in research, prioritizes the direction of progress over magnitude. Enjoys rigorous mental stimulation, relaxed social interaction, and regular outdoor exercise.
Miscellany
- Personal Interests: Nature, sunshine, the sea, and snowy mountains. Hobbies: Running, completed first marathon.
