🧭 AI/ML and System Architecture Article Map: From Theory to Practice

📚 Table of Contents¶

1. Introduction: Why You Need This Map
2. Station 1: Fundamentals and Core Questions (The Fundamentals)
3. Station 2: Deep Dives into Models and Generative AI (Deep Dives & GenAI)
4. Station 3: From Lab to Production (MLOps & System Architecture)
5. Station 4: Trends, Tools, and Real-World Cases (Future, Tools & Cases)
6. Closing: How to Use This Map

🌟 Introduction: Why You Need This Map¶

The scope of the field (Hook):
AI, ML, and system architecture are among the hottest and most complex domains in tech today — knowledge is dense and the pace of change is relentless.

The problem and the solution:
Insights are scattered across dozens of articles, making it hard to learn in a coherent sequence. This guide is your personal content map, helping you find the best starting point and reading path based on your current level and goals.

Who this is for:
- 🧩 Beginners: wanting to build a clear, structured knowledge base
- 🧠 Intermediate developers: wanting to go deep on specific algorithms or frameworks
- 🏗️ Senior architects: looking for production deployment experience to reference

🗺️ Station 1: Fundamentals and Core Questions (The Fundamentals)¶

Audience: Beginners, product managers, and non-technical readers who want to understand core concepts.
Core goal: Build a sense of the layers between AI, ML, and deep learning, and grasp common issues in the training process.

Introduction to AI and Machine Learning¶

1. Hung-yi Lee ML 2021: Basic Concepts of Machine Learning
2. Guide to Supervised and Unsupervised Learning (TBD)
3. Hung-yi Lee ML 2021: Training Techniques and Regularization
4. GenAI & ML Intro 2025 — Why Context Engineering Is the Key Technology in the AI Agent Era
5. GenAI & ML Intro 2025 — Deep Anatomy of Large Language Model Internals
6. GenAI & ML Intro 2025 — Technical Challenges and Pitfalls in Capability Benchmarking
7. GenAI & ML Intro 2025 — Understanding Machine Learning and Deep Learning
8. GenAI & ML Intro 2025 — From Pretraining to Human Value Alignment
9. GenAI & ML Intro 2025 — Lifelong Learning for General-Purpose Models
10. Generative AI & ML 2025 — Image and Audio Generation: The Dual-Track Convergence
11. AI Crossing the Rubicon — Self-Growth Journey (2026 Course Notes)

Model Evaluation and Tuning¶

1. Hung-yi Lee: Complete Guide to Model Evaluation Metrics (TBD)
2. Hung-yi Lee II: Solutions for Overfitting and Underfitting (TBD)
3. Hung-yi Lee: Data Preprocessing and Feature Engineering in Practice (TBD)
4. Hung-yi Lee ML 2021: Inside the Black Box — Explainable Machine Learning (XAI) Overview

🛠️ Station 2: Deep Dives into Models and Generative AI (Deep Dives & GenAI)¶

Audience: Developers with foundational knowledge who want to build and understand the principles.
Core goal: Go deep into mainstream model architectures and generative AI applications.

Mainstream Deep Learning and LLM Architectures¶

1. Hung-yi Lee ML 2021: Convolutional Neural Networks (CNN)
2. Hung-yi Lee: RNN and LSTM in Sequential Data Applications (TBD)
3. Hung-yi Lee ML 2021: Transformer and the Core LLM Architecture
4. Why Does Transformer Need Positional Encoding?
5. Course Recommendation — PyTorch Fundamentals: A Challenging but Worthwhile Course

Advanced and Specialized ML Applications¶

1. Hung-yi Lee ML 2021: Self-Attention Mechanism Principles
2. Hung-yi Lee ML 2021: GAN — Generative Adversarial Networks
3. Hung-yi Lee ML 2021: Autoencoder
4. Hung-yi Lee ML 2021: Self-Supervised Learning — BERT
5. Hung-yi Lee ML 2021: Handling Domain Shift — Domain Adaptation and Generalization
6. Hung-yi Lee ML 2021: Anomaly Detection
7. Hung-yi Lee ML 2021: Adversarial Attacks and Defense
8. Hung-yi Lee ML 2021: Reinforcement Learning — Core Concepts and Operational Framework
9. Hung-yi Lee ML 2021: Network Compression
10. Hung-yi Lee ML 2021: Lifelong Learning — How Machines Learn to Keep Learning
11. Hung-yi Lee ML 2021: Meta-Learning — Learning How to Learn
12. Hung-yi Lee ML 2023: Introduction to Diffusion Models

Large Language Model (LLM) Core Mechanisms and Applications¶

1. GenAI Era ML — LLM Training Tools in the Generative AI Age
2. GenAI Era ML — Technical Breakthroughs and Future Directions
3. GenAI Era ML — Dissecting the Internal Mechanics of Language Models
4. GenAI Era ML — Transformer's Competitors
5. GenAI Era ML — The Power and Limits of Pretrain-Alignment
6. GenAI Era ML — Post-Training, Catastrophic Forgetting, and Preserving Model Capability
7. GenAI Era ML — Large Language Model Reasoning Ability
8. GenAI Era ML — Challenges in LLM Evaluation and Efficient Inference
9. GenAI Era ML — Model Merging and Model Editing
10. GenAI Era ML — Beyond Text: Multimodal Dialogue and Generation
11. RAG Deep Dive: From Principles to Architectural Practice
12. RAG vs. Fine-Tuning: A Comparison
13. Fine-Tuning In Depth
14. LLaMA 3.2: Deployment, Testing, and Prompting
15. Gemini API: Getting Started and Applications
16. Accelerating Language Model Generation Speed — Li Hong-Yi 2026 Course
17. Harness Engineering (2026 Course Notes)

🏗️ Station 3: From Lab to Production (MLOps & System Architecture)¶

Audience: Senior developers and system architects.
Core goal: End-to-end practice from model deployment and monitoring to full system design.

MLOps Workflow and CI/CD¶

1. Introduction to MLOps and Workflow Design (TBD)
2. Deploying Model APIs with Docker and Kubernetes (TBD)
3. Getting Started with GitHub Actions Automation
4. From Command Line to Cloud: Deploying a Modern Python Service with gcloud CLI
5. LINE Bot in Practice: From Webhook to Cloud Run
6. LINE Bot Scheduled Push Notifications with Google Cloud Scheduler
7. TensorRT-LLM Learning Path — 12-Week Battle Plan

Cloud-Native and Containerized Architecture¶

1. Kubernetes Series I: Containerization Basics
2. Kubernetes Series II: Kubernetes Fundamentals
3. Kubernetes Series III: Advanced Topics and Deployment Strategies
4. Kubernetes Series IV: Enterprise Platform OpenShift
5. Microservice Architecture and ML System Integration (TBD)
6. Data Pipeline Design: From ETL to ELT (TBD)

🔭 Station 4: Trends, Tools, and Real-World Cases (Future, Tools & Cases)¶

Audience: Readers who want to follow the latest trends and see applied examples.
Core goal: Combine industry case studies with hands-on tools to spark ideas and applied thinking.

Agent Protocols and Integration¶

1. Prompting: Unlock Your AI — Introduction to the Model Context Protocol (MCP)
2. Understanding MCP — The Bridge Between AI and the Outside World
3. Building the AI Collaboration Future: MCP, A2A, and ACP — Roles and Applications
4. Image Generation vs. Stateful Tool Agent — A Comparison
5. Google AI Agent Development Kit (ADK) — Intensive Course Introduction

AI-Assisted Development Practices¶

1. AI Tools: A Guide to Rapid Development and Product Iteration
2. From Vibe Coding to Practice: How I Rebuilt My Personal Website with Google AI Tools
3. Vibe Coding Retrospective: From Writing Code to Navigating Code
4. Official vs. Community: Claude Code vs. Roo Code
5. Build Your Own AI Coding Assistant for Free with Google Colab
6. Python Toolbox: Automating Slide Decks with python-pptx
7. Website Automation: Hugo + GitHub Actions Static Site in Practice
8. From Prompt to Harness: As AI Giants Begin "Harvesting" Startups, Where Is the Engineer's Moat?
9. Refuse the Wild Horse — For Developers Wanting to Restart Manga Endings with AI: Context Engineering Practice Thoughts
10. When Coding Becomes Code Review — Senior Developer's AI Collaboration Survival Guide
11. From Execution to Problem Definition — The Evolution of AI Autonomy
12. When AI Gets Stuck in a Local-Optimum Deadlock: Why the Core Skill of the Future Isn't "Solving Problems" But "Defining the Problem Space"
13. Stop Being an AI Typist! Redefining Your Engineering Mindset with Specification-Driven Development with Coding Agents
14. Re-Reading "Continuous Delivery": From "What Abstract Nonsense" to a Pitfall Avoidance Guide in the AI Era
15. AI Lowers the Barrier to Development, But Raises the Barrier to Security
16. When AI Arrives "Today": Lessons from Lee Sedol's Move and the Paradigm Shift for Software Engineers

AI Applications and Trend Observations¶

1. LINE Bot Anti-Scam with Google Cloud — A Practical Project Walkthrough
2. Smart Traffic: Building a Violation Detection System with Computer Vision
3. Getting Started with Hardware AI Using ESP32
4. OpenClaw — Introduction and Positioning
5. When Masterpieces End Too Soon — Using AI as a Storyboard Editor to Revive Abandoned Stories
6. Large-Scale Web Architecture — When AI Floods the Toolbox: Future Architects Win by Defining Problems, Not Knowing Solutions
7. The Infinite Machine — A "Chicken Game" Over AI Control and a Survival Guide for the Rest of Us
8. Still Copying and Pasting for Your Financial Reviews? Use AI to Turn "Designing Your Financial Happiness" into Your Personal Butler
9. LINE Bot in Practice: Upgrading Your LINE Bot with AI and Rich Menus
10. Prediction Machines: Simple Economics of AI — A 2018 Prophecy for 2026

✅ Closing: How to Use This Map¶

• For beginners: Start at "Station 1: Fundamentals" to build a solid foundation.
• For architects: Jump straight to "Station 3: From Lab to Production" and focus on the MLOps and system design articles.