← Back to Blog

The AI Revolution: A Perspective from Experience

Tiago Leão • • 5 min read

After many years of building software across multiple industries — E-commerce, UK Retail, Betting & Gaming, and Investment Banking — I've witnessed numerous technological shifts. But none have been as transformative as what we're experiencing now with AI and Large Language Models.

The Shift is Real

This isn't like previous technology waves. Cloud computing changed where we ran code. Mobile changed how users accessed applications. But AI is changing how we write the code itself.

Every day, I use AI assistants to:

  • Generate boilerplate code in seconds
  • Debug complex issues by describing the problem
  • Refactor legacy code with natural language instructions
  • Write tests for edge cases I hadn't considered
  • Explain unfamiliar codebases quickly

The Productivity Multiplier

The numbers are hard to ignore. Tasks that took hours now take minutes. Not because AI writes perfect code — it doesn't — but because it handles the repetitive work while I focus on architecture, design, and the genuinely hard problems.

This isn't about replacing developers. It's about amplifying them.

Fine-Tuning: The Next Frontier

Generic AI models are powerful, but they don't know your codebase. They don't understand your architectural patterns, your naming conventions, or your business domain.

That's why I've been investing heavily in local LLM fine-tuning:

  • Privacy: Sensitive code never leaves my infrastructure
  • Customization: Models trained on my specific patterns and preferences
  • Speed: Local inference without API latency
  • Control: Full ownership of the models and their behavior

The Uncomfortable Truth

If you're in software engineering and you haven't started integrating AI into your workflow, you're falling behind. Not next year. Now.

This isn't gatekeeping or hype. It's observation. The engineers and teams I see thriving are the ones who've embraced these tools and are learning to use them effectively.

What This Means for Leaders

Technical leaders have a responsibility here:

  1. Model the behavior — Use AI tools yourself, visibly
  2. Create space for experimentation — Let teams try new tools
  3. Update expectations — Productivity baselines are shifting
  4. Address concerns honestly — Yes, some roles will change

The Path Forward

The engineers who thrive will be those who:

  • Learn prompt engineering as a core skill
  • Understand how LLMs work at a conceptual level
  • Know when to trust AI output and when to verify
  • Can fine-tune models for specific use cases
  • Stay curious and keep experimenting

My Approach

I practice what I preach. My home lab runs local models. I fine-tune on my own codebases. I use AI assistants daily while maintaining healthy skepticism about their outputs.

The goal isn't to blindly trust AI. It's to intelligently leverage it.

Want to discuss AI integration in software development? I find this topic endlessly fascinating. Get in touch.