AI EngineeringAvailable

RAG System Design

About this course

RAG is the dominant pattern for grounding LLM outputs in real data. This course covers chunking strategies, embedding models, vector databases, retrieval quality, prompt engineering for RAG, evaluation with RAGAS, and production patterns for cost and latency.

What you will learn

  • When RAG beats fine-tuning for knowledge-grounded answers
  • Chunking strategies: fixed, semantic, and recursive
  • Embedding models: OpenAI, Cohere, and open-source options
  • Vector databases: Pinecone, Weaviate, pgvector, and Qdrant
  • Retrieval quality: hybrid search, reranking, and MMR
  • Prompt engineering for RAG with citation
  • Evaluation with RAGAS: faithfulness and relevance
  • Production patterns: caching, streaming, and cost control

Your instructor

Ayodele Ajayi

Principal Engineer

Principal Engineer based in Kent, UK, with extensive experience across cloud-native security, platform engineering, and distributed systems. Ayodele has led engineering teams at scale and writes about what he learns — with a bias towards things that actually work in production.

Continue building your skills in this area.

MLOps Pipeline Engineering

8 lessons

LLM Fine-Tuning Fundamentals

8 lessons