Create your own AI chatbot with Mistral, Python, and Postgres

AI applications are increasingly important for driving innovation, improving customer service, and accelerating business decisions. This course introduces the practical development of an AI chatbot using Postgres, Python, and the Mistral LLM.

Designed for developers and architects, the course takes learners through the complete cycle of developing the data model, integrating a database with an externally hosted large language model, and combining conventional SQL queries with vector-based AI queries. By the end of the course, learners will have created a working proof of concept for each stage of the process and will be familiar with the technical requirements for building a successful AI chatbot.

• Quick recap of Postgres stored procedures, Python, and JSON
• Discussion of AI chatbot applications and their requirements
• Introduction to AI vectors, embeddings, and semantic search
• Interaction with large language model APIs, especially Mistral
• Retrieval augmented generation and prompt engineering
• The architecture of an intelligent chatbot
• Scraping a website and feeding a chatbot knowledge base
• Indexing for high-performance AI operations
• Using generative AI to create natural language results
• Making the chatbot more robust with data pipelines and the Model Context Protocol

On completion of this course, participants will be able to :

• Create a simple AI application that integrates databases and large language models (LLM)
• Understand the fundamentals of AI vectors, embeddings, nearest neighbor algorithms, and how to use them in Postgres
• Integrate a Postgres database with an LLM using the Python programming language
• Implement a semantic search engine that leverages the content of a relational database

• Create a chatbot that helps users navigate the content of a relational database, ask questions in natural language, and provide answers in natural language
  

The class is a 50/50 blend of lecture and hands-on labs. Students are each using their own PostgreSQL database, which they will use to define data models, to connect to a large language model, create a AI embeddings, and implement a working AI system. 

At the end of the class, students will have created a complete end-to-end working proof of concept that includes the database, the code to interact with the LLM, the queries needed for a semantic search engine, and the code for a fully functional chatbot that integrates with the database.

Students will be provided access to the Mistral LLM for the duration of the class to provide them with hands-on practical experience with AI and Large Language Models.