OpenAI
MIT
University of Cambridge
Université de Montréal
Rohan PandeyMember of Technical Staff · Periodic LabsResearcher↗
Researcher · Founder · Executive
Liam Fedus
Periodic LabsCo-founder
Fedus helped develop ChatGPT[1] and co-led GPT-4o's post-training, refining how it responds to users.[2] He co-created Switch Transformers, which scaled language models to a trillion parameters by activating only selected parts for each word. His research helped make larger models practical without a matching increase in computation.[3]
Career & education
Current rolePeriodic LabsSep 2025-presentCo-founder
Periodic LabsSep 2025-presentPreviously
Google Brain2020-2022Research Scientist
VP of Research2024-2025
Member of Technical Staff2022-2024
Google Brain2020-2022Work and education history is primarily focused on AI-relevant roles and may not be comprehensive.
Citation Trend
Total citations to scholarly work, including research papers, technical reports, books and theses, as recorded by Google Scholar. Each point is a dated snapshot.
citations87,301As of Oct 23, '26
citations+7,478Since last snapshot, Jun 15, '26
Selected contributions
GPT-4 Technical Report
With OpenAI, Joshua Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman and others ·Reported GPT-4's capabilities, limitations and safety evaluations as a text-and-image input model.
PaLM: Scaling Language Modeling with Pathways
With Aakanksha Chowdhery, Sharan Narang, Jacob Devlin, Maarten Bosma, Gaurav Mishra, Adam Roberts, Paul Barham and others ·Presented PaLM, a large language model demonstrating gains in reasoning, translation and coding.
Emergent Abilities of Large Language Models
With Jason Wei, Yi Tay, Rishi Bommasani, Colin Raffel, Barret Zoph, Sebastian Borgeaud, Dani Yogatama and others ·Examined abilities that appeared in larger language models but were absent in smaller ones.
References
- GPT-4o — language post-training leadsopenai.com
and co-led GPT-4o's post-training, refining how it responds to users.
- Switch Transformers — efficient trillion-parameter modelsarxiv.org
He co-created Switch Transformers, which scaled language models to a trillion parameters by activating only selected parts for each word. His research helped make larger models practical without a matching increase in computation.
Based on public sources. If accuracy is important, verify details using the links or references.
For corrections, updates, or requests to remove information, contact us.
Reviewed August 8, 2026 · Checked for accuracy and new information.
