8 Years in Tech: From React to AI & Blockchain Development

Hong Kong, 2018: Starting My Journey

It's 2018. I'm fresh out of university, starting my first job as a Junior Full Stack Developer in Hong Kong. React was just becoming mainstream, and I was diving headfirst into modern web development.

Eight years later, I'm writing this from Hong Kong, where I've built AI-powered applications serving 200K+ users, developed smart contracts processing 100K+ transactions, and evolved from a junior developer to an AI & Blockchain specialist.

This is that story.

Phase 1: The Foundation (2018-2019) — Learning Modern Web Development

React: Where It All Started

My first professional project wasn't glamorous. It was a responsive web application built with React and Node.js.

What I learned:

• How to build component-based UIs
• The power of state management
• That user experience matters more than technical perfection
• The importance of clean, maintainable code

First major lesson: Users don't care about your technology stack. They care if your application solves their problem efficiently.

Node.js: Full-Stack Capabilities

Learning Node.js opened up a new world of possibilities.

What this phase taught me:

• Full-stack development with a single language (JavaScript)
• RESTful API design and best practices
• Database integration with MongoDB and PostgreSQL
• Real-time features using WebSockets

Key realization: Modern web development is about building complete solutions, not just frontends or backends.

Phase 2: Scaling & Mastery (2019-2021) — Building Production Systems

Next.js: The Game Changer

2019: I discovered Next.js. It felt like React's missing pieces.

What clicked:

• Server-side rendering out of the box
• File-based routing (brilliant!)
• API routes (backend + frontend in one repo)
• Image optimization
• Performance by default

Projects I built:

• Scalable web applications serving 50K+ users
• Real-time features with WebSocket integration
• Payment gateway integrations
• Microservices architecture

What this phase taught me:

• Performance matters. A slow application loses users.
• Scalability is crucial. Design for growth from day one.
• TypeScript improves code quality. Type safety prevents bugs.
• Microservices enable flexibility. But they add complexity.

The TypeScript Transition

Around 2020, I fully embraced TypeScript.

Key decision: Invest in type safety, not just runtime speed.

Result: Fewer bugs, better IDE support, more maintainable codebases.

Phase 3: AI Integration Era (2021-2023) — Adding Intelligence

AI/ML: The Next Frontier

2021 was my year of AI exploration.

What I learned:

• Integrating OpenAI API and custom ML models
• Building AI-powered recommendation systems
• Natural language processing for user experiences
• Computer vision applications

First major AI project: An intelligent content recommendation system.

Challenge: Processing user behavior data in real-time.
Solution: Python ML models, Node.js API layer, Redis caching.
Result: 65% increase in user engagement.

Critical realization: AI isn't replacing developers—it's amplifying what we can build.

Python & Machine Learning

Learning Python opened doors to the AI/ML ecosystem.

Suddenly I could:

• Build custom ML models with TensorFlow
• Process large datasets efficiently
• Integrate AI services into web applications
• Create intelligent automation systems

Phase 4: Blockchain & Web3 (2023-2026) — The Decentralized Future

Google Cloud Platform: From VPS to Cloud

Before GCP: Shared hosting → VPS → Dedicated servers
After GCP: Cloud Run, Cloud Functions, Cloud SQL, Cloud Build

What changed:

• Infrastructure as code (Terraform)
• Automated deployments
• Global distribution by default
• Pay for what you use

Docker & Kubernetes: The Learning Curve

Week 1 of Kubernetes: "This is overcomplicated."
Week 4 of Kubernetes: "Okay, I see why this exists."
Month 3 of Kubernetes: "How did I live without this?"

Real-world use case at FMT:

• WordPress + MySQL on Kubernetes
• 1 writer replica, 3 reader replicas
• Auto-scaling during traffic spikes
• Zero downtime deployments

Result: Handled 260,000+ daily pageviews with 99.9% uptime.

Phase 5: Web3 Exploration (2021-2023) — The Detour

Smart Contracts & Solidity

In 2021, everyone was talking about Web3. I had to explore it.

What I built:

• NFT minting platforms
• Smart contracts for token transfers
• Blockchain explorers
• Integration with MetaMask

What I learned:

• Blockchain has legitimate use cases
• Most Web3 projects were vaporware
• The technology is fascinating
• The hype was exhausting

Decision: Stayed at the edges, didn't go all-in.

Why? My gut said: "This isn't mature enough yet. Real problems need solving now, not in Web3's eventual future."

Current take: Web3 has potential. But most users still need fast, reliable Web2 applications.

Phase 6: News/Media Mastery (2021-Present) — Finding My Niche

Free Malaysia Today: The 184% Growth Story

October 2021: Joined FMT as Technology Lead.

The situation:

• ~3M monthly users
• Older WordPress setup
• Limited performance optimization
• Solo operation (just me)

The challenge: Scale this platform while maintaining editorial workflow, integrating complex ad systems, and surviving breaking news traffic spikes.

The Transformation Journey

October 2024 → April 2025: The Big Migration

From:

• Next.js 12 (Pages Router)
• Monolithic WordPress
• Basic caching
• Manual deployments

To:

• Next.js 14.1 with ISR (Incremental Static Regeneration)
• Headless WordPress via GraphQL
• SmartCache system (my custom LRU cache with dependency tracking)
• Kubernetes-orchestrated infrastructure
• Cloudflare CDN with aggressive caching
• Zero downtime launch on April 18, 2025

The Numbers:

• Traffic growth: 3M → 8.5M monthly users (+184%)
• Load time reduction: 60% faster
• Uptime: 99.9%
• Memory optimization: 100% usage → 15% (saved the containers from crashing)
• Cache hit rate: 85-90%
• Database optimization: Query response time reduced from 800ms → 120ms

The SmartCache Crisis & Solution

The problem (February 2025):

• Memory usage hitting 100%
• Containers crashing during peak traffic
• 600,000+ posts overwhelming the cache
• Stale content showing to users

The solution I built:

// SmartNewsCache system with dependency tracking
// Every GraphQL query includes databaseId for surgical invalidation
// Size-based eviction (not just count)
// WebSub integration for real-time updates

Result: Memory dropped from 100% to 15%. Zero cache-related crashes since.

What made this work: Bias to action. I didn't wait for perfect requirements. I studied the problem, designed a solution, implemented it over a weekend, and deployed it Monday morning.

What Watching Traffic Grow Feels Like

Here's something I've never admitted publicly: It gives me immense pleasure when I see the traffic rise with each news story.

Why news/media is special:

• You publish an article → Traffic spikes within minutes
• Breaking news → Watch real-time analytics explode
• SEO improvements → See organic search climb daily
• Performance optimization → Users stick around longer

It's addictive. Like watching your code compile successfully, but it happens 24/7.

The Constants: What Hasn't Changed in 8 Years

1. Good Code Is Still Good Code

Whether it's React, Python, or Solidity:

• Readable > Clever
• Working > Perfect
• Maintainable > Optimized (unless you're at scale)
• Simple > Complex (until simplicity breaks)

2. Users Don't Care About Your Stack

Users care about:

• Does it work?
• Is it fast?
• Does it solve my problem?

Users don't care about:

• Your TypeScript configuration
• Your smart contract architecture
• Your AI model choice
• Your clever abstraction

3. Documentation Is Still Everyone's Weakness

Truth: Most developers (including me) hate writing documentation.

Reality: Future-you will hate past-you for not documenting that complex logic.

Solution: Write docs like you're explaining it to yourself in 6 months. Because you will be.

4. The Fundamentals Never Go Out of Style

After 8 years and multiple technology shifts, here's what still matters:

• Algorithms & Data Structures (essential for AI and blockchain)
• Database Design (whether SQL or blockchain state)
• HTTP & REST (even Web3 uses similar patterns)
• Git (learned in 2018, still using it daily)
• Problem Solving (the meta-skill that transcends all tech)

5. Continuous Learning Isn't Optional

2018: React, Node.js
2019: Next.js, TypeScript
2021: Python, AI/ML
2023: Solidity, Web3
2026: Still learning (currently exploring advanced AI models and cross-chain protocols)

At 29 years old, I'm constantly learning new technologies. Why? Because in AI and blockchain, the landscape changes every few months. Comfort kills careers.

The Unexpected Lessons

Lesson #1: AI & Blockchain Require Different Mindsets

Building AI applications taught me:

• Data quality matters more than model complexity
• User experience is crucial - AI features must feel natural
• Iteration is key - ML models improve with feedback
• Ethics matter - AI decisions affect real people

Building blockchain applications taught me:

• Security is non-negotiable - Smart contract bugs can't be patched
• Gas optimization is an art - Every operation costs money
• Decentralization has trade-offs - Not always the right solution
• User education is essential - Web3 UX is still terrible

Combining both: The most exciting projects combine AI intelligence with blockchain trust.

Lesson #2: StackOverflow Made Me Better

Top 3% contributor (14,111 reputation). Not because I'm smarter than others, but because:

Teaching forces clarity.

When you answer someone's question, you can't hand-wave. You must:

• Understand the problem deeply
• Explain it simply
• Provide working code
• Anticipate edge cases

Result: Answering 694 questions made me a better engineer than reading 1,000 tutorials.

Lesson #3: Metrics Drive Decisions

Everything I've built has numbers attached:

• 65% engagement increase with AI features
• 200K+ monthly active users
• 100K+ blockchain transactions processed
• 50% faster AI inference response times

Why? Because "I improved the app" means nothing. But "I increased engagement by 65% with AI recommendations" is quantifiable. Measurable. Provable.

In interviews, in promotions, in salary negotiations—numbers win.

Lesson #4: Technology Choices Are Bets, Not Marriages

Bad mindset: "I'm a React developer."
Good mindset: "I solve problems. Currently using React, but I'll use whatever works best."

Outcome:

• When React wasn't enough → Learned Next.js
• When JavaScript wasn't suitable → Learned Python for AI
• When traditional databases weren't right → Learned blockchain state management

Career advice: Fall in love with problems, not tools. The best solution might require learning something new.

Lesson #5: Side Projects Accelerate Learning

Building personal projects taught me:

• Freedom to experiment - No client constraints
• Learning by doing - Building DApps taught me more than reading docs
• Portfolio building - Real projects > tutorial projects
• Problem discovery - Personal projects reveal real challenges

My side projects:

• AI-powered personal assistant
• NFT marketplace
• DeFi yield farming calculator
• Blockchain-based voting system

Life lesson: Side projects aren't distractions—they're learning accelerators. Build things that interest you.

The Hard Truths (Things I Wish Someone Told 22-Year-Old Me)

1. Impostor Syndrome Never Fully Goes Away

8 years in, built applications serving 200K+ users, developed smart contracts processing 100K+ transactions.

And still: Every time I start a new project, there's a voice saying "Everyone here knows more than you."

Truth: That voice lies. But also, it keeps you learning. Double-edged sword.

2. Your First Job's Technology Doesn't Define Your Career

Started with React in 2018. If I'd believed "I'm a React developer forever," I'd be missing out on AI and blockchain opportunities.

The industry moves fast. Your first tech stack is training wheels. Learn the fundamentals, then adapt.

3. Salary Negotiations Get Easier With Proof

Early career: "I'm worth X because... I need it?"
Mid career: "I'm worth X because I built Y serving Z users with A% improvement."

Data wins. Build portfolio evidence, not just years of experience.

4. Specialization Matters, But So Does Versatility

In AI and blockchain, deep expertise matters. But being able to work across the full stack is also valuable.

My approach: Deep knowledge in AI/Blockchain, broad knowledge in full-stack development.

Result: I can lead projects end-to-end, from smart contracts to AI features to frontend UX.

5. Work-Life Balance Is Essential (I Learned This Early)

20s: Worked long hours, but also made time for learning and side projects.
Late 20s: Realized sustainable pace > burnout sprints.

Current philosophy: 40-hour productive weeks with continuous learning > 80-hour exhausted weeks.

What I'd Tell 22-Year-Old Jaycee

Learn TypeScript Earlier

Seriously. I added TypeScript in 2020. Should've been 2019.

Why it matters: Type safety catches 70% of bugs before production. Time saved debugging > time spent writing types.

Start Writing Technical Blogs Sooner

Regret: I waited until 2026 to start blogging consistently.

Lost opportunity: 8 years of experience could've been 50+ blog posts, establishing thought leadership.

Advice: Document your learning journey. Your "today obvious" is someone's "tomorrow revelation."

Contribute to Open Source More

GitHub is great. But OSS contributions show:

• You can work in teams
• You understand large codebases
• You write production-quality code

Current plan: Start contributing to React, Next.js, or blockchain ecosystem projects.

Build a Portfolio From Day One

Reality: My portfolio went live in 2026. Should've started in 2018.

Why? GitHub stars and testimonials compound over time.

Lesson: Your work has value. Showcase it.

Ask for Help Earlier

Stubborn truth: I always try to solve everything alone first.

Result: Wasted weeks on problems that someone else solved in 10 minutes.

New approach: Google for 30 minutes. If stuck, ask in Discord/Slack/GitHub. Ego isn't worth the time loss.

Where I Am Today (January 2026)

The Numbers

• Age: 29 years old
• Experience: 8+ years professional
• Current role: AI Full Stack & Blockchain Developer
• Location: Hong Kong
• GitHub: Active contributor to open source projects
• Built: Applications serving 200K+ users
• Technologies: React, Next.js, Node.js, TypeScript, Python, AI/ML, Solidity, Web3, Smart Contracts

What I'm Looking For Next

I'm excited about the intersection of AI and blockchain.

Ideal projects:

• AI-powered blockchain applications
• Decentralized AI platforms
• Smart contracts with AI decision-making
• Full-stack applications with AI integration
• Innovative Web3 solutions

What I bring:

• Proven track record (200K+ users, 100K+ transactions)
• AI & Blockchain specialist
• Full-stack expertise
• 8 years of continuous technology evolution
• Bias to action, metrics-driven, innovative thinking

What I need:

• Challenging problems (easy is boring)
• Innovative projects (cutting-edge tech)
• Results-driven culture
• Competitive compensation

Final Thoughts: The 8-Year Meta-Lesson

Here's what 8 years taught me that no bootcamp or tutorial ever could:

Technology is just the tool. Problem-solving is the craft.

• React → Next.js → AI/ML → Blockchain
• All just tools
• The craft is understanding:
  • What problem needs solving?
  • Who experiences this problem?
  • What's the simplest solution?
  • How do we measure success?

Your career is a marathon, not a sprint.

• Sprint: Work 80 hours/week for 2 years, burn out
• Marathon: Work sustainable pace for 8+ years, still going strong

Keep learning, or become obsolete.

• The technology you know today will be legacy in 5 years
• The fundamentals you learn today are forever
• Balance both

Document everything.

• Your future self will thank you
• Your teammates will thank you
• Your interviewers will thank you
• Your blog readers will thank you

Find what makes you excited to open your laptop.

For me: Building AI-powered applications and smart contracts that solve real problems.

For you: Could be anything. Find it. Double down on it.

Thank You, 2018

Starting my career in 2018 with React was the beginning of an incredible journey.

8 years later, I'm still proving it.

• To myself
• To my employers
• To my users (200K+ of them)
• To anyone who doubted that a junior developer could become an AI & Blockchain specialist

The technologies changed.

React → Next.js → Python → Solidity → AI/ML → Web3 → ?

The fundamentals didn't.

Problem-solving. Continuous learning. Bias to action. Metrics-driven decisions. Building solutions that matter.

And the journey continues.

At 29, I'm exploring advanced AI models and cross-chain protocols. Not because I need to. Because I can. Because learning is what keeps you alive in this industry.

Here's to the next 8 years.

Let's Connect

Want to discuss AI integration, blockchain development, or full-stack architecture?

I'm available for:

• Technical consulting
• Full-time opportunities (Hong Kong or remote)
• Coffee chats about AI & blockchain projects
• Portfolio reviews

Find me:

• Portfolio: jaycee-tran-portfolio.vercel.app
• GitHub: Check out my code
• Email: jaycee.tran1996@gmail.com
• Discord: jaycee_tran1996
• Telegram: @jayceetran1996

Or just:

• Schedule a call
• Drop me an email
• Comment below

P.S. If you're a hiring manager looking for an AI Full Stack & Blockchain Developer who can build intelligent, scalable solutions—let's talk.

Found this useful? Share it with someone starting their tech journey. Or someone 8 years in and wondering what's next. Spoiler: The best is yet to come.