Skip to main content
Free community · Weekly live sessions · Open-source curriculum

AI Engineering you can actually ship to production

Practical, production-grade AI systems engineering, taught through 63 open-source notebooks, weekly free live sessions, and a hands-on engineering community.

For AI engineers, ML engineers, backend engineers transitioning into AI, MLOps engineers, and technical founders building AI products.

63 open-source notebooks | 750+ engineers in community | Weekly live sessions | Free to join
Community engineers from Microsoft Capgemini Databricks Goldman Sachs Startups Worldwide

AI & ML Engineers

Already building models. Want to get them into production reliably—eval, serving, observability, the works.

Backend Engineers → AI

Strong SWE foundations. Want to own RAG pipelines, agent architectures, and LLMOps instead of just calling the API.

MLOps Engineers

Managing infra for ML. Want to extend that to LLM systems—prompt versioning, drift detection, cost management.

Technical Founders

Building AI products. Need to make the right architectural calls early before the technical debt compounds.

Open-source · Free on GitHub

RAG Engineering — 63-Notebook Curriculum

13 stages, 63 runnable notebooks, all built on a single shared corpus — so every technique is directly comparable. Naive RAG through enterprise production. No paywalls. No signups.

View on GitHub
00

Evaluation Harness

Cross-notebook P@k baseline. Run after changing indexes or fusion.

01–04

Foundations

Naive RAG, chunking strategies, hybrid retrieval (FAISS + BM25 + RRF), reranking.

05–07

Query Intelligence

Query rewriting, multi-query expansion, HyDE (Hypothetical Document Embeddings).

08–09

Core Retrieval

Graph RAG, agentic retrieval with tool-using agents.

10–12

Models & Index

ColBERT late interaction, embedding benchmarking, quantized indexes.

13–15

Advanced Chunking

Contextual retrieval, proposition chunking, parent-child chunking.

16–19

Evaluation

RAGAS, DeepEval, SPLADE sparse retrieval, MMR diversity.

20–23

Generation-time RAG

FLARE, Self-RAG, LLMLingua compression, context ordering.

24–26

Deployment

Qdrant production setup, semantic caching, distributed tracing.

27–34

Production Engineering

Async pipelines, streaming, batch processing, failover, access control, audit logging, cost monitoring, data governance.

35–42

Advanced Topics

Multimodal RAG, temporal reasoning, federated retrieval, causal reasoning.

43–46

Enterprise Projects

Domain-specific RAG for legal, medical, financial, and technical documentation.

47–62

Production Apps & Infra

Customer support, code search, research assistants, incident response, scaling, security, compliance, privacy.

All notebooks share the same Orion Analytics corpus — so every technique is directly comparable. Runnable, free, no signups.

Open the Curriculum →

How it works

Three entry points. All free to start.

WhatsApp Community

700+ engineers discussing RAG, agents, MLOps, and production architecture daily. Active discussions, shared notebooks, live Q&A.

Join free →

Weekly Live Sessions

Weekly live engineering sessions announced in the WhatsApp group. Code-first, no slides, real system walkthroughs with live Q&A. Join the group to get session links.

Join WhatsApp for session links →

Open-Source Notebooks

63 runnable notebooks on GitHub. Foundations → query intelligence → ColBERT → evaluation → FLARE → production engineering → enterprise projects.

Browse curriculum →

What the curriculum actually covers

Naive RAG Hybrid Retrieval HyDE Graph RAG ColBERT SPLADE RAGAS DeepEval FLARE Self-RAG LLMLingua Qdrant Semantic Cache Multimodal RAG Federated RAG

From the Engineering Blog

Production AI systems. No fluff, no hype—just implementation.

Free · No credit card · No catch

Come build in public with us.

700+ engineers in the WhatsApp community working through RAG, agents, MLOps, and production AI together. Weekly live sessions, shared notebooks, real architecture discussions.

If you're building AI systems and want to be around engineers doing the same—this is where that happens.

Join the Free WhatsApp Community