Beyond the Paper: Real-World AI in Action

The Social Artificial Intelligence Night (SAINT) of the Department of Computer Science & Security is an important meeting point for scientists, industry partners, and enthusiasts in the field of artificial intelligence in Austria to exchange information about the latest advancements and projects.

Keynote Announcement: 

Loubna Ben Allal, Machine Learning Engineer, Hugging Face (🤗)

Loubna Ben Allal, is a machine‑learning engineer at Hugging Face, where she leads the SmolLM project — an open‑source effort to build compact, high‑performing large language models. In addition to directing SmolLM, she is a core contributor to the BigCode project and co‑author of the The Stack dataset and the StarCoder family of code‑generation models.

Loubna holds two master’s degrees—in Mathematics and Deep Learning — from École des Mines de Nancy and ENS Paris‑Saclay. After completing two internships, she joined Hugging Face in 2022, initially contributing to language models for code generation. Her work quickly expanded to broader LLM research, positioning her as an expert on model efficiency, open‑source tooling, and responsible AI development. Since joining Hugging Face, Loubna has driven the design, training, and deployment of several generations of SmolLM (including SmolLM 2 and the more recent SmolLM 3). Her recent “SmolLM Training Playbook” — presented as a comprehensive, book‑style blog post — offers a deep dive into the practical and methodological decisions that enable efficient training of large language models at scale.

Loubna’s blend of deep technical expertise, open‑source advocacy, and clear communication makes her a compelling voice on the future of compact, accessible large language models. In her keynote, she will be discussing the latest advances in efficient LLM training and the lessons learned from designing and training SmolLM 3.

The Smol Training Playbook.

Thanks to our partners: