Large Language Models for Devs: from zero to your first LLM application
summary
Shows a concrete developer path: assemble AAC pipelines that index documents, run BM25 and embedding retrievers (sentence-transformers) and inject top docs into generators to stop hallucinations — the Italian Super Cup question returns the correct winner via RAG. Demos use Haystack, Ollama and hosted OpenAI models, plus local quantized models like Mistral 7B and Llama variants that run on a laptop. Pipelines serialize to YAML and serve as FastAPI endpoints; evaluation and deployment tips cover retriever metrics, observability and hybrid BM25/vector setups.
AI-written summary · about
chapters (19)
- 0:00Introduction to LLMs for devs
- 1:11Why this talk on LLMs
- 2:16How LLMs are trained
- 5:09Hands-on with generators
- 7:44Prompt engineering basics
- 8:30Challenges with LLMs
- 9:13Retrieval augmented generation
- 12:05Keyword vs vector retrieval
- 14:45Embeddings and vector search
- 17:58Tips for production apps
- 21:47Evaluating LLM pipelines
- 22:59Deploying with Haystack
- 24:06Multilingual use cases
- 25:10Information extraction demos
- 25:56Generating quizzes from URLs
- 27:22Observability in production
- 27:47Function calling for apps
- 28:11Agents and future trends
- 29:43Q&A session
related talks
30mLearning to wield large language models
27mAI Demystified Unveiling the Intelligence Behind the Machine
31mDeploying Large Language Models with Ease
01h 34mA Practical Starter's Guide to building LLM based projects
01h 49mAI engineering with open access LLMs that lie, curse, and steal
18mLLMs and GenAI: A Guide to Building Your Own RAG Model for Document Interaction
01h 19mBuilding an LLM on your Laptop to...
40mDeploying Your Own Language Models: A Use Case Demo with Open Source Tools