LLM Development Course
3.4k 2026-04-30
iusztinpaul/hands-on-llms
A hands-on course to learn LLMs, LLMOps, and vector databases by building, training, and deploying a real-time financial advisor LLM system.
Core Features
End-to-end development of LLM systems, including training, real-time data streaming, and inference pipelines.
Practical experience with LLM fine-tuning using QLoRA and proprietary datasets.
Integration of vector databases (Qdrant) for RAG and real-time data ingestion (Alpaca, Bytewax).
Implementation of LLMOps practices with Comet ML for experiment tracking, model registry, and monitoring.
Deployment of LLM pipelines on serverless GPU infrastructure (Beam) and AWS.
Detailed Introduction
This project offers a comprehensive, hands-on educational course designed to teach the intricacies of Large Language Models (LLMs), LLMOps, and vector databases. Participants learn by actively building a real-time financial advisor LLM system, covering the entire lifecycle from data ingestion and model training (using QLoRA) to real-time inference and deployment. It emphasizes practical application of tools like Qdrant, Comet ML, Beam, and LangChain, providing a robust understanding of how to design, implement, and manage production-ready LLM solutions.