Talks | Manuel Faysse

Selected talks

Pretraining small language models & CroissantLLM

IBM Research Paris · Meta FAIR Paris (AI Alliance) · Naver Labs France · Paris GenAI School · DGFIP & DINUM · Digital Innovation Makers Summit, Station F · Ateliers de l’IA, Illuin Technology · BackProp

Watch at BackProp

Document retrieval in vision space & ColPali

Unbabel · LlamaIndex · Jina AI Retrieval Session at EMNLP · Crédit Agricole Investment Banking · Amazon Research

Watch at LlamaIndex

Contextual embeddings

Meta FAIR Paris

Press & interviews

Zeta Alpha: ColPali: document retrieval with vision-language models — podcast
MIT Technology Review: ‘Copyright traps’ could tell writers if an AI has scraped their work
Nature: Has your paper been used to train an AI model? Almost certainly
Usine Digitale: CroissantLLM, un modèle d’IA open source et bilingue
Usine Nouvelle: CroissantLLM: interview & coverage

ColPali also features in State of AI 2024, Thoughtworks Technology Radar (Oct. 2024), and DeepLearning.AI’s Multi-Vector Image Retrieval course.

Teaching & mentoring

ENSAE & CentraleSupélec. Teaching assistant for Pierre Colombo’s NLP courses.
Paris GenAI School, 2024. Lecture on language model pretraining.
EPFL. Teaching assistant in physics, mathematics, programming, and applied machine learning.

Guest lectures on NLP and ML privacy; supervision of research internships, student projects, UNESCO Tech for Good initiatives, and health ML hackathons.