Build a Multimodal GenAI App with LangChain and Gemini LLMs
In this 90 minute hands-on workshop, we'll learn how to build a multimodal GenAI app that recommends products from scratch. By the end of the workshop, you'll have built a clone of Fashion Buddy, an application that lets people search for apparel recommendations based on an uploaded photo.
We build the app in Python and use RAG (retrieval augmented generation) with LangChain, Google’s Gemini LLM (large language model), and vector search with Astra DB.
Session Agenda
Throughout the 90-minute workshop, all attendees will learn:
- What is RAG and how does it work with LangChain and vector search?
- How to ingest + vectorize data from CSV file into Astra DB
- How Google Gemini interprets images
- How to use LangChain for ingestion, embedding, and prompt retrieval
- How to deploy a user interface with Streamlit to take image uploads and return similar results
Handy Links:
Fashion Buddy GitHub Repo
Speakers
Melissa Herrera
Developer Relations
DataStax
Cedrick Lunven
Software Engineer
DataStax