From Zero to Full-Stack: How I Built My Learning Path?

Last winter, I was squatting in my rented room staring at the screen—my product Flash Warehouse had been live for three days, and the user count was zero. At that moment, Nietzsche's 'become who you are' was echoing in my mind, but the code wasn't running, and philosophy couldn't save me.

Just a few months earlier, I was working as a product manager at a company, drawing prototypes and writing requirement docs every day. Suddenly I decided to quit my job and start my own business, writing a warehouse management system from scratch. Many people asked me: you studied computer science, how did you suddenly become a full-stack developer?

TL;DR I started learning programming in my freshman year, from Java backend to Vue frontend, then Three.js and IoT—every step was forced by real projects. There are no shortcuts, only constant pitfalls and retrospectives.

How did I go from only writing Java backend to learning frontend?

In my sophomore year, I took on a freelance project and found that I had finished the backend but no one was doing the frontend. That was my first 'full-stack' experience—forced.

At that time, I only knew Java and Spring Boot; I couldn't even remember HTML tags. The client needed an admin panel, so I gritted my teeth, downloaded layui, and copied code from the docs. Looking back, that page was ugly enough to make you cry, but at least it worked.

Later I realized that frontend wasn't a question of 'to learn or not to learn' but 'when you are forced to learn'. During my internship at Loctek, I had to modify the frontend of factory control systems; at Yonshine, I had to use Three.js for digital twin dashboards. Every time I switched projects, I had to learn a new framework.

It's like playing a game—you never know what skills the next boss will require, but you have to be ready to upgrade your skill tree. According to the Stack Overflow 2024 Developer Survey, full-stack developer is the most popular role ^[1], but not many can truly handle a project solo.

How did I 'cross over' into IoT and hardware programming?

When I was working on the digital twin dashboard at Yonshine, I needed to display device data in real-time on 3D models. That was my first encounter with MQTT and Raspberry Pi.

Honestly, I had no concept of hardware at first. STM32, ESP32, serial communication—these terms sounded like gibberish. But the project couldn't wait, so I spent two weeks cramming an 'Embedded Linux Basics' book, then deployed an edge computing node on a Raspberry Pi 4B, cutting hardware costs by 30% ^[2].

Later, I applied this experience to Flash Warehouse—supporting IoT devices like barcode scanners and Bluetooth printers. Once you've stepped on this landmine, you understand: the boundary between hardware and software is actually blurry; the key is understanding how data flows.

What skills are needed to build a SaaS product solo?

After starting my own business, I realized that full-stack isn't just about the tech stack; it's also about product, design, and operations.

Although Flash Warehouse is written by me alone, the tech stack includes: Spring Boot backend, Vue 3 frontend, uni-app mobile, Next.js website, Docker deployment, Cloudflare Workers for CDN. The GitHub 2024 report says AI-assisted coding tools can improve efficiency by 30-55% ^[3], and I deeply feel that—Cursor saved me a lot of time writing boilerplate code.

But technology is just the tip of the iceberg. I also have to design database schemas, draw UI prototypes, write user docs, and do SEO. Sometimes I feel like I'm not a developer but a 'one-person company'.

This table looks glamorous, but behind every skill are countless nights debugging until dawn.

What is the biggest pitfall in learning full-stack?

The biggest pitfall is not technical difficulty, but trying to learn everything and mastering nothing.

When I first started learning full-stack, I wanted to learn all frameworks in a month. One day React, next day Vue, then Flutter. The result was that I could only write 'Hello World' in each, and would be exposed in an interview.

Later I learned my lesson: project-centric learning. I only learn the technology I need for the project at hand. For digital twin, I dived deep into Three.js; for SaaS, I focused on Spring Boot and Vue. It's like weight training—progressive overload: each time you lift a little more than last time, but never stop.

According to IDC, the penetration of cloud-native technology in SMEs grew from 15% in 2020 to 45% in 2025 ^[4]. This means full-stack developers need to continuously learn new things, but the key is to build a knowledge system, not chase novelties.

Three suggestions for those who want to get into full-stack

If you ask me now: how should a beginner start full-stack? I would say: first build a complete project, even if it's terrible.

First, choose a main direction. Backend or frontend? Master one first, then expand. I myself mastered Java backend first, then filled in frontend later.

Second, learn by doing projects. Don't just read books; build a blog system, a warehouse management system, anything. The project I did for the university competition, though the code was terrible, gave me my first taste of the full development process.

Third, accept imperfection. Your first project will be terrible, and the second won't be much better. But as Sisyphus rolls the boulder—every time you push it up, you become stronger.

Key takeaways

• Full-stack is forced by projects, not learned from books
• Learn project-centrically, don't chase new frameworks
• Technology is just the tip of the iceberg; product thinking and ops are equally important
• Accept imperfection, finish first then perfect

Now I'm still on the entrepreneurial path, iterating Flash Warehouse. But every time I write a line of code, I remember the sophomore me staring at the layui docs blankly. Back then I couldn't even write frontend well, but at least I dared to write.

---

References

1. Stack Overflow 2024 Developer Survey — Full-stack developer is the most popular role
2. Raspberry Pi Official Site — Raspberry Pi 4B used for edge computing
3. GitHub Octoverse 2024 — AI-assisted coding tools improve efficiency by 30-55%
4. IDC Cloud Native Report — Cloud-native penetration in SMEs grew from 15% to 45%