Modern Problems Require Modern Solutions - Finding Your Meme Twin with Embeddings and Vector Databases

Learn how to build a meme twin finder application using embeddings and vector databases in this conference talk from NDC Oslo 2025. Discover the fundamentals of embeddings and how they transform unstructured data like text and images into hyper-dimensional arrays called vectors. Explore vector databases, understanding their purpose and functionality for storing and searching embeddings efficiently. Follow along as the speaker demonstrates a practical application that converts popular memes—from Ancient Aliens Guy to Zombie Boy—into embeddings stored in a vector database. Examine the code implementation and database queries that power the meme twin detection system. Participate in live demonstrations where audience photos are processed to find matching meme characters, combining entertainment with technical education about modern data processing techniques.