This is my eighth year as a frontend developer. It's not a year-end review, but I wanted to use this milestone to reflect: what does a frontend engineer's growth path actually look like, and what things truly matter?
My Timeline
2017: Junior frontend. jQuery, CSS, the occasional Vue project
2018: Started this blog. Webpack, Babel, ES6+ became daily tools
2019: React and TypeScript became my main stack. Started caring about engineering
2020: Working from home due to COVID. Lots of learning time. Early Vue 3 and Vite adopter
2021: First time independently owning a complex frontend module
2022: Started giving tech talks. Deep dives into Next.js and React architecture
2023: The year AI tools went mainstream. Started thinking about "what an engineer really is"
2024: Taking on more technical decisions. Building a design system
2025: Officially became tech lead. Balancing technology and management
2026: Now
What Actually Matters
1. Fundamentals beat frameworks
I wrote a lot about Vue and Webpack in 2018, but all of that has since changed. What hasn't changed is:
- HTTP: caching, request optimization, status codes
- Browser rendering: CRP, reflows and repaints, compositing layers
- JavaScript fundamentals: prototype chains, Event Loop, closures
- Algorithms and data structures: not every hard problem, but solid foundations
- Design patterns: no need to memorize, but recognize them when you see them
Framework knowledge has an expiration date. Fundamentals don't.
2. Writing is the best way to learn
Over these eight years, I've written 100+ blog posts. Every article forced me to turn vague understanding into clear expression. Many things I "thought I understood" — I only realized I didn't when I tried to write about them.
If you can only do one thing to improve yourself, make it writing.
3. Communication skills are rarer than technical skills
In the first three years of your career, technical ability is the main variable. After that, communication and influence become increasingly important.
Being able to explain technical decisions to non-technical people, getting your team to buy into your proposals, earning genuine respect in code reviews — these are harder than writing elegant code, and worth more.
4. Choose the "hard" option
Every time I make a technical choice, I ask: which option will teach me more?
Of course, project requirements can't be ignored, but when conditions allow, actively choose options with learning value. Don't always stay in your comfort zone.
My Long-Term Take on AI Tools
I started using AI tools seriously in 2023 — about two to three years ago now. A few observations:
AI has increased the speed of "writing code," but hasn't reduced the need for "thinking about design"
Requirements analysis, architecture design, trade-off decisions — AI still doesn't handle these well. Maybe not because AI isn't smart enough, but because these judgments depend heavily on context: company culture, team capability, historical technical debt, and so on.
AI has made "average quality" easier to achieve
Anyone using AI can write mid-tier code. So where does the difference between engineers lie? I think it's in: the ability to identify problems (knowing what to solve) and the ability to judge quality (knowing what a good solution looks like).
Conclusion: Keep learning, don't be anxious
Technology changes, but the core value of engineers hasn't: understand the problem, design the solution, deliver the result.
What I'd Tell My Past Self
If I could go back to 2018, I'd tell the me back then:
- Don't chase every new framework; pick one and go deep
- Write articles; put what you learn into words
- Work on your English — good English opens many doors in docs and community
- Contribute to open source, even if it's just filing an issue
- Work alongside people who are better than you — it's the fastest way to grow
What I'm Doing Now
This year I started contributing to an open-source React component library (still in early stages). I've given a few tech talks. I've mentored two solid engineers on my team.
No grand ambitions — just want to keep being a reliable engineer.
April 8, 2026. Eight years.