Maxime Labonne is a Senior Staff Machine Learning Scientist at Liquid AI, where he leads the post-training division focused on refining large language models for production use. He holds a PhD in Machine Learning from the Polytechnic Institute of Paris and is recognised as a Google Developer Expert in AI and Machine Learning, with a research portfolio that spans graph neural networks, model fine-tuning, and the evaluation methods that determine whether an LLM is genuinely useful in enterprise contexts.

Before Liquid AI, Labonne held senior machine learning roles at JP Morgan Chase & Co and Airbus, where he applied research-grade modelling to fraud detection, document understanding, and applied industrial problems. That combination of regulated-industry experience and frontier research is part of what makes his teaching unusually credible for engineering audiences. He has made significant open-source contributions, including the widely used LLM AutoEval and LazyMergekit tools and frontier models such as NeuralBeagle and Phixtral.

Labonne is the author of two best-selling technical books published by Packt, the LLM Engineer's Handbook and Hands-On Graph Neural Networks Using Python. Both have become reference texts for practitioners building production AI systems, and both are part of the ExpertEdge book to course catalogue, transformed into structured video, modular reading and assessments designed for enterprise LMS deployment.

His current focus bridges research and application, with particular emphasis on scalable architectures, inference optimisation, and the responsible deployment of language models at enterprise scale. He is widely followed for his educational work, including the GitHub-hosted LLM Course that has become a starting point for many practitioners moving into LLM engineering. For L&D leaders running technical upskilling programmes, his books on ExpertEdge offer the kind of depth senior engineers actually engage with, the topic explored in our analysis of why your engineering team isn't engaging with the LMS.