A Reference Learning Plan
This is not a rigid curriculum to follow strictly, but a learning roadmap tailored to the needs of CS students in the AI era. Everyone's background, goals, and schedule differ — adapt it to your own situation.
Design Principles
Foundation first, AI acceleration second. Traditional CS fundamentals (data structures, algorithms, operating systems, computer networks) remain hard currency in the AI era. AI tools can accelerate your learning but cannot replace your understanding of the basics.
Balance theory and practice. Every course recommends video lectures (theory) + labs/projects (practice). Watching videos without doing projects means you'll forget everything.
AI tools throughout. From freshman year, integrate AI-assisted programming, prompt engineering, and similar tools into your daily learning — don't treat them as isolated courses.
Phase 1: Freshman Year — Building Foundations
Goal: Build programming and math foundations, establish good study habits.
Fall Semester
| Topic | Recommended Resources |
|---|---|
| Python basics | Python Official Tutorial |
| C programming | Harvard CS50 |
| Calculus I | MIT 18.01 |
| Linear algebra | 3Blue1Brown Essence of Linear Algebra + MIT 18.06 |
Practice: Solve 1-2 easy LeetCode problems daily. Use AI as an assistant, not as an answer machine.
Spring Semester
| Topic | Recommended Resources |
|---|---|
| Data structures & algorithms | UC Berkeley CS61B |
| Calculus II | MIT 18.02 |
| Probability & statistics | MIT 6.041 |
| Git & GitHub | Git section + GitHub Tutorial |
Practice: Complete all CS61B labs and projects. Start using Git to manage your code repositories.
Summer recommendations: - Study Harvard CS50 for consolidation - Build small Python projects (web scraper, data analysis, simple game) - Get initial exposure to AI programming tools (Copilot, ChatGPT)
Phase 2: Sophomore Year — Core CS + AI Introduction
Goal: Master core CS courses and start exploring AI-related topics.
Fall Semester
| Topic | Recommended Resources |
|---|---|
| Computer architecture | UC Berkeley CS61C |
| Operating systems | MIT 6.S081 |
| Introduction to ML | Andrew Ng Machine Learning |
| AI-assisted programming | AI Coding section |
Practice: Complete CS61C RISC-V labs and MIT 6.S081 xv6 labs. Start using Claude Code or Cursor in real projects.
Spring Semester
| Topic | Recommended Resources |
|---|---|
| Computer networks | Stanford CS144 |
| Database systems | CMU 15-445 |
| Deep learning | Andrew Ng Deep Learning Specialization |
| Prompt engineering | Prompt Engineering section |
Practice: Implement a B+ tree index for a simple database; complete an image classification project with PyTorch.
Summer recommendations: - Find a summer internship or contribute to open source - Go deep on an AI application area (NLP/CV/Agents) - Learn Docker and basic DevOps skills
Phase 3: Junior Year — AI Specialization + Career Prep
Goal: Specialize in your chosen direction while preparing for graduate school or job market.
Fall Semester
| Topic | Recommended Resources |
|---|---|
| Distributed systems | MIT 6.824 |
| Reinforcement learning | David Silver RL Course |
| RAG systems | RAG section |
| Agent development | Agent section |
Practice: Complete a full AI project (RAG knowledge base, multi-agent system). Start preparing for entrance exams if applicable.
Spring Semester
| Topic | Recommended Resources |
|---|---|
| LLM application engineering | AI Engineering section |
| AI safety | AI Safety section |
| Compilers (elective) | Stanford CS143 or Crafting Interpreters |
Practice: Build a portfolio (GitHub repos + project demos). Entrance exam prep at full intensity. Job seekers start Leetcode grinding and interview prep.
Phase 4: Senior Year — Capstone and Career
Goal: Complete capstone project, determine career direction.
- Graduate school track: Prepare for interviews, contact advisors early, read papers in your target area
- Job track: System design interview prep, continue deepening your tech stack
- Study abroad track: Prepare application materials, accumulate research experience
About the CS Entrance Exam (408)
If you plan to take the Chinese CS graduate entrance exam, these four subjects each account for roughly 25 points:
| Subject | Points | Recommended Course |
|---|---|---|
| Data Structures | ~25 | UC Berkeley CS61B |
| Computer Architecture | ~25 | UC Berkeley CS61C |
| Operating Systems | ~25 | MIT 6.S081 |
| Computer Networks | ~25 | Stanford CS144 |
After completing the above courses, use domestic exam prep materials (e.g., Wangdao series) for targeted review. International courses help you understand principles; exam prep materials help you pass the exam — the combination works best.
Principles for Using AI Tools
- AI is an amplifier, not a replacement. AI helps you write code faster, but responsibility for understanding the code remains yours.
- Understand first, then use. Don't use AI to generate code you don't understand. When you encounter something you don't grasp, ask AI to explain the principle.
- Use AI to learn, not to do. Have AI explain concepts, translate documents, and summarize papers — but thinking and decision-making must be your own.
- Keep your coding hands sharp. Even with AI tools, regularly write code by hand to maintain sensitivity to syntax and algorithms.
This plan is based on the author's personal learning experience and discussions with multiple professors and senior students. Since the author's focus is on applied practice, coverage of theoretical CS and mathematics is more limited. If you have suggestions, please reach out via email.