Track · 52 lessons · 27.2 h
AI Engineering
Build apps on top of LLMs, from prompts and RAG to agents and evals.
0 / 52 done
LLM basics4 lessons · 1h
Retrieval & RAG10 lessons · 4h 45m
- What is RAG and When Does It Beat Pure Promptingeasy15 min
- Embeddings Explained: From Text to Vectorseasy15 min
- FAISS Basics: A Vector Search You Can Run Locallymedium25 min
- Vector Databases Beyond FAISS: Metadata Filtering at Scaleeasy15 min
- Choosing an Embedding Model (with a Local Fallback)medium30 min
- Chunking Strategies That Matter More Than Your Embedding Modelmedium50 min
- Hybrid Search: Combining BM25 Keyword + Vector Retrievalmedium15 min
- Reranking Retrieved Results with Cross-Encodersmedium15 min
- Build a RAG Pipeline from Scratchmedium1h
- Confidence Gates: When Your RAG Should Refuse to Answermedium45 min
Evaluation3 lessons · 1h 50m
Agents8 lessons · 4h 55m
- What Is an Agent and When Do You Actually Need Onemedium25 min
- The ReAct Loop: Thought → Action → Observationhard55 min
- Designing a Tool Registry for Agentsmedium35 min
- Adding MCP Tool-Use to Your RAG Assistanthard50 min
- Agent Memory: Managing the Context Windowmedium30 min
- Reflection & Self-Correction with an LLM Judgemedium35 min
- Error Handling in Autonomous Agentsmedium30 min
- Building Retrieval Tools for Agents (Web + Internal KB)medium35 min
Safety & security5 lessons · 2h 45m
- Safety Guardrails: Regex + LLM-Judge Patternsmedium35 min
- Defending Against Prompt Injectionmedium35 min
- PII Redaction & Output Sanitizationmedium30 min
- RBAC for RAG: Chunk-Level Access Controlmedium35 min
- Audit Logging for Compliance (SIEM-Ready)medium30 min
Data & fine-tuning11 lessons · 6h 30m
- Building a RAG Dataset Pipeline End-to-Endhard55 min
- Ingesting Public Corpora (SQuAD → Your Schema)medium30 min
- Relevance Filtering: Keeping Your Corpus On-Topicmedium30 min
- Deduplicating Datasets with MinHashmedium35 min
- Quality-Filtering a Training Datasetmedium30 min
- Schema Validation & Atomic Writes for AI Datasetsmedium30 min
- Train/Val/Test Splitting Without Leakagemedium30 min
- Synthesizing Training Data: Refusals & Multi-Turnmedium35 min
- Formatting Data for Supervised Fine-Tuningmedium30 min
- LoRA Fine-Tuning as a Fallbackhard55 min
- Publishing Models: Adapter vs Merged to the Hubmedium30 min
Production3 lessons · 2h 30m
Interview prep8 lessons · 2h 54m
- How to Explain a RAG Architecture in an Interviewmedium25 min
- Common Interviewer Questions on Embeddings, Retrieval, Fine-Tuningmedium25 min
- How to Defend Evaluation Metrics and Admit Failure Modesmedium20 min
- Resume Bullets for AI Projects: What to Write, What to Dropeasy12 min
- How to Explain an Agent Architecture in an Interviewmedium25 min
- Common Interviewer Questions on Agents & Tool Usemedium25 min
- System Design: Enterprise RAG Walkthroughhard30 min
- Resume Bullets for Agentic & Tool-Use Projectseasy12 min