← All Courses LIVE

Build a Production-Ready RAG System from Scratch

Agentic RAG System

Phase 1 of The Mother of AI (MOAI) Project. You'll build a production-grade Agentic RAG system end-to-end — an arxiv paper curator powered by Airflow pipelines, hybrid retrieval (BM25 + vectors with RRF), LLM generation via Ollama, and an agentic layer for query validation, document grading, and adaptive retrieval. Full observability with Langfuse, Redis caching, streaming APIs, and multiple client interfaces. No toy demos — 23+ tools, real architecture, real code you can run and ship.

12 hours Self-paced 29 lessons
Enroll Now — $99

Substack Annual Premium subscribers get 50% off — Contact us

Agentic RAG System

What You'll Learn

7 weeks. From zero infrastructure to a production RAG system with observability.

Infrastructure & Data Engineering

The foundation real systems depend on.

Docker-orchestrated services, FastAPI, OpenSearch, PostgreSQL, and Airflow. Production habits from day one — service boundaries, retries, failure handling, and health checks.

Data Ingestion & Real-World Pipelines

Where most quality gains come from.

Automated pipelines that fetch, parse, and normalize academic PDFs. Messy inputs, unreliable APIs, parsing failures — the problems that dominate real systems.

Retrieval as a First-Class Problem

Not an afterthought — the core of RAG.

BM25 keyword search, semantic embeddings, structured chunking, and hybrid ranking. Keyword and semantic signals combined deliberately for stable, explainable retrieval.

Observable, Agentic Production System

Measure and iterate, not guess.

LLMs as a controlled component, not the centerpiece. Streaming, citations, full tracing, and caching. Then agentic workflows with LangGraph — query validation, document grading, and adaptive retrieval.

Tools & Technologies

23+ production tools you'll master

One Project — 23 Modern AI Engineering tools organized by layer: Application & Development, Data Pipeline, Retrieval & Storage, Inference, and Observability

Docker-based service orchestration

FastAPI with health checks & docs

PostgreSQL for structured storage

OpenSearch setup

Airflow integration

Local LLM with Ollama

All course code is open source

Complete notebooks and production-ready Python scripts — browse the code before you enroll.

View on GitHub

Learn from experienced practitioners — 18+ years building and deploying AI/ML systems at real companies.

How It Works

Self-paced curriculum that fits your schedule. Live support when you need it.

Self-Paced Recorded Curriculum

Six weeks of production-focused lessons. Learn at your own pace, on your own schedule. All labs are in Python scripts, runnable and production-ready.

Discord Support

Join our private Discord community for participants. Get help from peers, share progress, and collaborate daily.

Completion Certificate

A formal certificate recognizing your achievement, suitable for L&D budgets and professional development.

Lifetime Access

Retain access to all course materials, recordings, and future updates forever.

6-8 Hours Per Week

Build real systems alongside your full-time job. Designed to fit your busy schedule.

What you need

Prerequisites

Make sure you have these covered before starting.

Comfortable with Python (intermediate level)

YouTube

Basic understanding of APIs and HTTP requests

YouTubeDocs

Familiarity with Git and the command line

YouTube

Basic Docker knowledge helpful but not required

YouTubeCompose

No prior ML/AI experience required — we start from foundations

Ready to Build?

Get lifetime access to 12 hours of hands-on content, real code, and production patterns.

Enroll Now — $99

Substack Annual Premium subscribers get 50% off — Contact us

Course Curriculum

8 modules · 29 lessons · 12 hours of content · Free Preview Available

01

Kickoff Session

1 lessons

Welcome session to kick off the course, set expectations, and walk through the roadmap.

  • Kickoff Session Recording from 16th Nov 2025
02

Week 1 — The Infrastructure That Powers RAG Systems

4 lessons

Set up the production infrastructure foundation: Docker, FastAPI, PostgreSQL, OpenSearch, Airflow, and Ollama.

  • Pre-Read (Must read before watching)
  • Introduction to RAG
  • Understanding the Architecture & Project Structure
    Free Preview
  • Infrastructure Walkthrough & Question & Answers
03

Week 2 — Bringing Your RAG System to Life: The Data Pipeline

5 lessons

Build automated pipelines that fetch, parse, and normalize academic PDFs with real-world error handling.

  • Pre-Read (Must read before watching)
  • Introduction to Data Pipeline & Ingestion in RAG
  • Question & Answers Part 1
  • Ingestion Walkthrough
  • Question & Answers Part 2
04

Week 3 — The Search Foundation Every RAG System Needs

4 lessons

Implement keyword search with BM25, OpenSearch index design, and build a debuggable retrieval flow.

  • Pre-Read (Must read before watching)
  • Understanding the Basics of Search (BM25)
    Free Preview
  • Code Walkthrough & OpenSearch Keyword Retrieval
  • Question & Answers
05

Week 4 — Chunking Strategies & Hybrid RAG System

5 lessons

Explore chunking strategies, generate embeddings, and combine BM25 + vector search with Reciprocal Rank Fusion.

  • Pre-Read (Must read before watching)
  • Introduction to Chunking Strategies
  • Question & Answers Part 1
  • Code Walkthrough
  • Question & Answers Part 2
06

Week 5 — The Complete RAG System

4 lessons

Wire hybrid retrieval to LLM generation with streaming APIs, optimized prompts, and a chat interface.

  • Pre-Read + File Downloads (Must read before watching)
  • Architecture Revisit
  • Code Walkthrough
  • Question & Answers
07

Week 6 — Monitoring & Caching

4 lessons

Add end-to-end RAG tracing, latency and error monitoring, Redis caching, and production reliability patterns.

  • Pre-Read + File Downloads (Must read before watching)
  • LLM Observability Overview
  • Question & Answers
  • Code Walkthrough
08

Week 7 — Agentic RAG (Bonus)

2 lessons

Build agentic workflows with LangGraph — query validation, document grading, and adaptive retrieval.

  • Pre-Read (Must read before watching)
  • Agentic RAG with Code Walkthrough

Learn from people who've actually built these systems

This isn't theory from a textbook. Every module comes from real experience building and deploying RAG systems in production. We've made the mistakes so you don't have to.

We've intentionally priced this course low because we want people to actually take advantage of it and learn. Put your money in the right place — invest in skills that compound, taught by engineers who ship.

Frequently Asked Questions

Our own experience is the biggest differentiator. We've both built and deployed AI/ML systems in production serving millions of users every day, with 18+ years of combined experience. This course is experience-driven - built over months and years in production settings, not from a weekend demo project. The recordings reflect real production insights that very few are aware of. Check our profiles before making your decision. Shirin's LinkedIn Shantanu's LinkedIn
RAG is fundamentally an information-retrieval problem. By starting here, you learn search, ranking, retrieval, and system design - skills that naturally extend to LLMs, Agents, and recommendations. Read our blog on this
Yes. We plan to add ColPali/VLM-based retrieval and Graph RAG modules to this course. When they are ready, they will be added here - and since you have lifetime access, you get them at no extra cost.
This course is hosted on Payhip. After purchasing, you'll receive an email with login details to access all course materials through your Payhip account. View on Payhip Login to Payhip
Premium Annual subscribers get a 50% course discount, access to advanced blogs with video walkthroughs, and monthly live calls. They also get early access to new courses and deeper technical discussions. Subscribe to Premium
This course is fully recorded from a real cohort run. You get all code walkthroughs, debugging sessions, and architectural discussions on demand.
There is no fixed cohort start date. You get immediate access and can move at your own pace.
Around 6-8 hours per week is sufficient to follow along. You can go faster or slower depending on your schedule.
That's completely fine. The 7-week plan is a guideline - you have lifetime access and can complete it at your own pace.
Yes. The course teaches transferable architecture and patterns that apply across cloud, on-prem, and different vendors.
Yes. The bonus Week 7 covers the foundations of Agents through Agentic RAG - including query validation, document grading, multi-step reasoning, and LangGraph orchestration. It's a bonus module included with the course.
Yes. Many learners expense it as professional development, and invoices can be provided if needed.
No. The focus is on where real systems break: data pipelines, retrieval quality, evaluation, reliability, and observability.
This course is not for people looking only for prompt tricks or shortcuts. It's designed for those who want to build real AI systems end to end.
AI systems are increasingly retrieval- and agent-driven. Starting now gives you a strong foundation before this becomes the default expectation.

Ready to Build?

Get lifetime access to 12 hours of hands-on content, real code, and production patterns.

Enroll Now — $99 View All Courses

Substack Annual Premium subscribers get 50% off — Contact us