OPEN LEARNING CURRICULUM · BEGINNER TO ADVANCED

Understand agents.
Design systems.
Build what matters.

A hands-on learning platform spanning Agentic AI, RAG, system design, GPU computing, CUDA, and production LLM inference—from first principles to end-to-end architecture.

Explore the curriculum →Explore the repository ↗

09 learning modules

Build-first practical approach

Python + Mermaid implementation & architecture

START HERE · MODULE 01

Build your foundation.

Learn what an agent is, how its architecture fits together, and how the execution loop turns a goal into a sequence of actions.

01 / FOUNDATIONS OF AGENTIC AI

From a language model to a working agent.

Start with first principles. Understand the model, tools, planner, memory, state, and orchestrator. Then trace an agent through observe, reason, plan, act, and evaluate.

Agent architectureExecution loopTools & memoryPythonInterview prep

Open Module 1 guide →

AGENT EXECUTION LOOPConceptual architecture

flowchart TD
  A(["OBSERVE
Goal + context"]) --> B["REASON
Understand current state"]
  B --> C["PLAN
Choose next action"]
  C --> D{"ACT
Tool needed?"}
  D -->|Yes| E["Validate + authorize"]
  E --> F["Execute tool"]
  F --> G["EVALUATE
Inspect result"]
  G --> H{"Goal complete?"}
  H -->|No| A
  H -->|Yes| I(["FINAL RESPONSE"])
  D -->|No| I

THE CURRICULUM

Nine modules. One learning path.

Move from core concepts to production-grade system design. Each module connects concepts to implementation and engineering trade-offs.

MODULE 01Start here

Foundations of Agentic AI

Agents, models, tools, memory, planning, execution loops, and a first Python agent.

Open learning guide →

MODULE 02Learning guide

RAG Fundamentals

Why retrieval-augmented generation exists, its architecture, and the basic RAG pipeline.

Explore module →

MODULE 03Guide + project

Document Processing & Embeddings

Ingestion, parsing, chunking strategies, embeddings, and indexing decisions. Includes a Python pipeline with sample documents and tests.

Explore module guide →Open Python project ↗

MODULE 04Detailed guide

Information Retrieval

Lexical search, BM25, dense retrieval, vector search, and retrieval evaluation.

Explore module →

MODULE 05Detailed guide

Advanced Retrieval

Hybrid search, reciprocal rank fusion, reranking, query rewriting, and expansion.

Explore module →

MODULE 06Detailed guide

Advanced RAG

HyDE, contextual compression, Self-RAG, CRAG, Adaptive RAG, and Graph RAG.

Explore module →

MODULE 07Detailed guide

Agent Architecture & Orchestration

Tool orchestration, multi-agent patterns, state, memory, and coordination.

Explore module →

MODULE 08Detailed guide

Production Engineering

Evaluation, observability, reliability, security, latency, and cost controls.

Explore module →

MODULE 09Detailed guide

End-to-End System Design

Bring the pieces together through complete system design projects and trade-offs.

Explore module →

Content status: All nine modules now have expanded learning guides with Mermaid architecture diagrams, worked examples, design trade-offs, and practice. Modules 1–3 also include original architecture illustrations; Module 3 includes a hands-on Python project.

HOW TO LEARN

Learn it. Build it. Explain it.

A repeatable study rhythm that turns concepts into practical engineering skills.

STEP 01

Understand the concept

Start from first principles. Learn the problem, vocabulary, and architecture.

STEP 02

Trace the system

Follow data and control flow with diagrams, examples, and failure scenarios.

STEP 03

Implement in Python

Build a small working version. Add tests and inspect its behavior.

STEP 04

Defend trade-offs

Practice design questions around scale, cost, latency, reliability, and safety.

SPECIALIZED LEARNING TRACK · GPU-01 → GPU-36

GPU computing & AI infrastructure.

Go beneath AI applications: understand GPU hardware, learn CUDA, map deep-learning operations to parallel execution, and design efficient LLM inference systems.

BEGINNER → ADVANCED

From computer architecture to production inference.

Start with CPU/GPU fundamentals, memory hierarchy, SMs and warps. Progress through CUDA kernels, tensor computation, prefill/decode, KV cache, inference engines, multi-GPU serving, and production optimization.

GPU architectureCUDAPyTorchLLM inferencevLLMMulti-GPU

Start GPU-01 → Full GPU curriculum ↗

GPU LEARNING PATHConceptual progression

flowchart TD
 A["Computer architecture"] --> B["GPU hardware & memory"]
 B --> C["Parallel programming"]
 C --> D["CUDA kernels"]
 D --> E["Deep-learning computation"]
 E --> F["LLM inference architecture"]
 F --> G["Inference engines"]
 G --> H["Multi-GPU production systems"]

GPU-01Start here