What You’ll Learn

• Build a RAG-powered retriever for restaurants: Encode restaurant data with WatsonX Slate embeddings and store them in Chroma for efficient similarity search.
• Create and coordinate multi-agent systems with CrewAI: Define agents like the Profile Builder, Coarse RAG Recommender, Food Trend Analyst, and Recommendation Finalizer to work collaboratively.
• Analyze multimodal user data: Use vision-capable LLMs to extract insights from dining photos and integrate them with textual reviews.
• Incorporate real-time web search: Use Serper to fetch trending cuisines and new restaurants, ensuring recommendations stay current.
• Deliver structured outputs with Pydantic: Standardize agent outputs for reliability and seamless downstream processing.
• Run and orchestrate an end-to-end pipeline: From user history to final ranked recommendations, you’ll deploy a fully functional AI restaurant recommender.
• Build a Gradio Web App for online chatbot!

Who Should Enroll

• Early-career data scientists and ML engineers who want hands-on experience building multi-agent systems.
• AI enthusiasts curious about CrewAI, RAG, and multimodal LLM applications.
• Students and researchers interested in applying cutting-edge AI to real-world recommendation challenges.
• Developers in food-tech or hospitality who want to explore AI-driven personalization.

Why Enroll

What You’ll Need

• Basic Python programming knowledge.
• Familiarity with LLM concepts and vector databases (helpful but not required).
• (Optional) Sign up for a free Serper API account as a part of the project for real-time online search.
• Interest in AI agents, recommender systems, and multimodal data.