IDUG NA 2025

Large language models (LLMs) have revolutionized AI by enabling a new category of applications called language apps. Enterprises have widely started using these apps for various purposes such as chatbots, document summarization, question–answering, and information retrieval. This has sparked renewed interest in AI in recent years. A crucial aspect of language apps is the generation of vector embeddings using language models. This process converts unstructured content, like PDFs, into AI–readable representations. These embeddings are used for vector similarity search, which offers more precise semantic search compared to traditional methods. However, vectors have high dimensionality, so they need cost–effective and secure storage, as well as efficient search methods. Db2 is introducing its first phase of vector support, which includes in–database vector storage, vector similarity search, and various vector operations. In this presentation, the Speaker will discuss Db2’s upcoming vector support features and how they can be utilized to develop modern LLM applications with Db2.