Curriculum v1 — A project-based roadmap covering agentic AI, RAG, retrieval engineering, orchestration, production systems, and end-to-end system design.
Learn the fundamentals. Build the systems. Understand the trade-offs.
A responsive GitHub Pages learning site is included in this repository: Agentic AI Systems — Learn. Design. Build..
To publish it, open Settings → Pages in the repository and set the build/deployment source to GitHub Actions. The included workflow deploys the static site on pushes to master. The first deployment may take a few minutes; check the Actions tab for its status.
Every topic includes: fundamentals, Mermaid architecture diagrams, real-world examples, selected research papers, Python implementation with tests, module-wise interview questions, company-style system design questions, explained answers, practical exercises, and a knowledge check.
See CURRICULUM.md for the complete 9-module roadmap. The full Module 1 learning guide includes concepts, architecture diagrams, sequence diagrams, failure handling, a system design exercise, research papers, and interview answers.
The repository has a guide page for each module. All nine modules now have expanded learning guides with Mermaid diagrams, worked examples, implementation patterns, tests or testable exercises, system-design trade-offs, interview walkthroughs, and practice with answers. Modules 1–3 also include original architecture SVG illustrations. Module 3 has a runnable Python document-processing and embedding project with sample data and unit tests.
- Module 1 — Foundations of Agentic AI — full learning guide
- Module 2 — RAG Fundamentals
- Module 3 — Document Processing and Embeddings
- Module 4 — Information Retrieval
- Module 5 — Advanced Retrieval
- Module 6 — Advanced RAG
- Module 7 — Agent Architecture and Orchestration
- Module 8 — Production Engineering
- Module 9 — End-to-End System Design
- Module 1 Python implementation and tests
- Module 3 document-processing and embedding pipeline — sample Markdown/text ingestion, chunking, metadata, optional embeddings, JSONL output, semantic search, and tests.
This is an independent case-study and practice track, separate from the nine-module learning curriculum.
- Company Engineering Practices — public engineering patterns and primary sources.
- Hands-on Company Case Studies — six complete exercises with Mermaid diagrams, Python implementations, unit tests, evaluation plans, and trade-offs.
- Company-Style System Design Interview Practice — original company-inspired design prompts, not verified interview questions.
- Module-wise interview question bank — beginner, intermediate, and advanced questions for all 9 modules.
- Company-style system design practice — representative prompts inspired by engineering problem areas at Google, Meta, Amazon, Microsoft, OpenAI, Anthropic, and NVIDIA. These are practice prompts, not claims about exact questions asked in company interviews.
Follow the learning journey and AI research explainers on Instagram: @frontierai.daily
CURRICULUM.md— modules and topicsMODULE_1_FOUNDATIONS.mdthroughMODULE_9_*.md— module guides with embedded Mermaid diagramsCOMPANY_PRACTICES.md— independent company engineering case-study librarycompany-practices/— six runnable company-inspired case studiesSYSTEM_DESIGN_INTERVIEWS.md— company-inspired system design practice promptsinterview-preparation/MODULE_WISE_QUESTIONS.md— interview questions organized by modulePROGRESS.md— learning trackerresearch-papers/— paper links and reading notesinterview-preparation/— question banks and answersimplementations/— Python projectscontent/— Frontier AI carousel and Reel planning
Review commits for personal information, API keys, credentials, private datasets, and copyrighted PDFs. Prefer linking to official paper pages or open-access versions instead of uploading copyrighted PDFs.
A dedicated beginner-to-advanced track covering computer architecture, GPU execution and memory, CUDA programming, deep-learning computation, LLM inference, serving engines, multi-GPU systems, and production operations.
- Complete GPU curriculum — 36-topic structured roadmap.
- GPU-01: Computer Architecture for GPU Programmers — first chapter with Mermaid diagrams, worked examples, exercises, calculations, and answers.
The GPU track complements the Agentic AI modules and System Design resources. It is designed to teach both the underlying hardware concepts and how those concepts affect real AI inference systems.