8-Month Full-Stack + AI Learning Roadmap for CS Students

Overall Learning Path

The recommended sequence is: Programming Basics → Frontend Development → Java Backend → Frontend-Backend Integration → AI Application Development → Algorithms (by interest) . The target is to independently build a management system with AI features, explain code, debug issues, and deploy it. Estimated duration: ~8 months at 1–2 hours daily.

Six Stages with Time Estimates and Deliverables

Stage 1: Development Basics (2 weeks) – Files & directories, command line, Git, program execution, browser-server model, HTTP, JSON. Outcome: run projects, commit/rollback changes, understand a request lifecycle.

Stage 2: Web Fundamentals (6 weeks) – HTML, CSS, Flexbox; JavaScript variables, functions, arrays, objects, events, async requests. Outcome: build a worker list with search, add, delete.

Stage 3: Vue 3 Frontend (4 weeks) – Vue 3, components, reactivity, form validation, routing, API calls, state management; then TypeScript. Outcome: produce a management page with pagination, edit modal, loading and error states.

Stage 4: Java & Database (6 weeks) – Java syntax, OOP, collections, exceptions; SQL, joins, keys, indexes, transactions. Outcome: design simple business tables, query and maintain data with Java.

Stage 5: Backend & Integration (6 weeks) – Spring Boot, API design, layered architecture, MyBatis, auth, logging, testing, deployment. Outcome: complete a management module with real frontend-backend data flow.

Stage 6: AI Application (6 weeks) – Model APIs, prompt engineering, structured output, streaming, knowledge retrieval (RAG), evaluation; Python as needed. Outcome: add an AI assistant with cited sources to the labor system.

Each stage extends the same project, making later code easier to comprehend.

Frontend Mastery in Three Steps

Static Pages : HTML/CSS for worker list, search box, forms, detail page.

Page Interaction : JavaScript for filtering, form validation, array CRUD; understand Promise and async/await.

Backend Connection : Vue for pagination, data persistence, file upload, error handling.

Before Vue, ensure you can manipulate an object array for search, edit, delete. Mini-programs are common now; learn them after JS and API calls are solid, then add WeChat-specific lifecycle, login, upload, location.

Backend Learning Around a Complete Business Flow

Start from project development. Focus on three responsibilities:

Controller : receive request, validate input.

Service : enforce business rules (e.g., duplicate job type rejected).

Mapper / Data Access : query and update database.

Gradually add pagination, parameter validation, unified error handling, authentication, transactions, tests. SQL must cover CRUD, joins, grouping, and understanding why indexes speed queries. Recommended references: Dev.java official tutorial for Java basics; Spring official REST example for API practice.

AI Split into Two Tracks

Track 1: AI Application Development

Core skills: call models, design I/O, connect business systems, retrieve documents, evaluate answers. Example projects: smart customer service, document Q&A, complaint summarization, info extraction.

Track 2: AI Algorithms & Models

Core skills: data processing, math foundations, ML, neural networks, training & fine-tuning. Example projects: classification, image recognition, model training & optimization.

Start with AI Application Development integrated into the project. Step-by-step exercises:

Call a model to summarize a complaint.

Require fixed fields: issue type, key facts, items to verify.

Stream response in frontend, handle timeout/failure.

Add RAG: retrieve relevant docs, let model answer with citations.

Prepare test set; check correctness, grounding, and graceful "insufficient info" responses.

Existing Java backend can call model APIs; Python can be learned later for data processing and algorithm work. For deeper algorithm track, sequence: Python, NumPy/Pandas → linear algebra, probability, calculus → ML & evaluation → PyTorch → Transformer & fine-tuning. References: PyTorch tutorials and Hugging Face Chinese course.

Practice Routine

Daily: 30 min reading, 60 min coding, 15 min logging issues. When using AI assistance, first ask it to explain the call flow and give hints, then modify and verify code yourself. Aim to articulate why each line is written. Solid fundamentals combined with AI tooling yield high efficiency.