Loading...
AI Engineering course · one course, beginner to advanced
Data engineering for ML, retrieval and RAG in production, and running agents in production: pipelines, ingestion, versioning and data contracts, then chunking, retrieval and grounding, then tool calls, planning, memory and failure handling.
All 206 lessons plus every future lesson, with runnable code and enterprise diagrams. One payment, lifetime access.
The retrieval half of every LLM product, taught where it actually breaks: the scale cliff between a demo and 5 million documents, chunking, reranking, hybrid search, and the document ingestion layer that quietly caps everything downstream.
The data foundation every model depends on: batch and streaming ingestion, ETL and ELT, lakes and warehouses, validation, versioning, labeling, imbalanced data, synthetic data, and data contracts.
Measured on real shop data: what a feature is worth, point-in-time joins, freshness, windows, online and offline consistency, late events, encodings, missing values, versioning, cost, embeddings, and a real feature store.
Measured on the features chapter's model: what a model file holds, saving formats, library version skew, pinned dependencies, container images, reproducible artifacts, a real model registry, signatures, ONNX conversion and the model supply chain.
Measured on a quiet machine with the course's own model: batch or online, where request time goes, tail latency, batching, workers and threads, cold starts, honest load tests, caching, timeouts and fallbacks, many models on one box, and cost per prediction.
Putting an LLM in a loop, done reliably: why per-step errors compound, the tool-calling contract, durable execution, memory and context, and why constraining an agent beats maximizing its autonomy.
206 lessons in total, across 3 levels
Where the course starts: how production ML works (packaging, serving, pipelines, feature stores, monitoring), then how language models read text, with tokens, tokenizers, embeddings and search by meaning. Start here.
74 lessons7 chapters
Evaluating LLM systems, operating them in production, and securing them: error analysis, judges and eval statistics, fine-tuning vs RAG, inference and guardrails, then prompt injection and agent safety.
62 lessons3 chapters