In recent times, large language models (LLMs) have been a subject of significant interest in the natural language processing (NLP) community, mainly due to their remarkable performance across a wide spectrum of language-related tasks. In this presentation, my focus will be on how we have utilized LLMs in our "Devoxx Insights" application, specifically to provide talk summaries, semantic search, Q&A, and photo search. I will provide numerous code examples to illustrate how to utilize GPT in your own projects and how to deploy them effectively into production. By the end of this talk, attendees will gain a better understanding of the potential and capabilities of LLMs and how to leverage them for their own NLP projects.