At Sansan Global Development Center (SGDC), the team I manage, Haraya, closes every stand-up meeting with a phrase that guides our work: *’change lives.’* It’s more than just words; it’s a reminder of why we do what we do.
For nearly three decades, I’ve worn many hats: software engineer, team lead, architect and engineering manager. My journey started in the days of DOS-based applications like Clipper, FoxPro, and C++. Back then, every change meant waiting for the whole program to compile just to discover if an error had slipped in. Fast-forward to today, and the landscape looks entirely different: IntelliSense, real-time error feedback, reference navigation, and even live coding. Development is faster, smarter, and far more collaborative.
But unlike many managers my age who have moved away from coding to focus solely on people and project management, I’ve stayed curious. I wanted to see if I could still perform at my best as both a leader and developer.
The image shows Team Haraya displaying their phrase, “Change Lives.”.
Returning to the Craft
At SGDC, I was fortunate to be part of an organization that listens. When I shared my desire to explore beyond management, I was given the opportunity to develop a mobile application using React Native. I had dabbled in it before, but this time I wanted my approach to be more AI-driven.
To be honest, it was both exciting and intimidating. Some colleagues even asked, “Nag develop pa diay ka, sir?” which translates to, “Oh, you’re still coding, sir?” Their surprise motivated me, but it also came with pressure. I had to prove, to them and to myself, that I could still deliver.
My AI-Powered Development Journey
Most of the team leaned on tools like Devin, Claude, and Cursor. I chose GitHub Copilot because of its seamless integration with VS Code. My first tasks were ambitious:
- Build a POC – OCR reader.
- Create an entry form that saves data into SQLite
With AI’s help, I built the OCR reader in just two days. It wasn’t perfectly architected, but it successfully scanned and generated text. The entry form was ready in three days.
Without AI, the OCR reader alone would have easily taken me a month, thousands of lines of code, endless library research, and countless Stack Overflow searches. The contrast was astonishing.
The image shows a POC project I initiated using Copilot to generate a Method of Contact form, modifying the selection input from buttons to radio buttons. The AI automatically produced the corresponding code implementation, integrating my existing TextInput component.
What AI Really Changed
AI didn’t just accelerate my work it changed what I focused on. Instead of obsessing over indentation, unused declarations, code conventions, or boilerplate code, I could spend more time on what really matters:
- Redesigning code structure
- Ensuring solutions are maintainable, reusable, and scalable
- Keeping performance and security at the core
By offloading repetitive coding tasks to AI, I was free to think like an architect again, not just a coder.
Human-in-the-loop
AI is an incredible accelerator. It can generate code snippets, automate routine tasks, and shorten development timelines dramatically. But speed doesn’t always mean quality.
The real difference between fragile, short-lived software and systems that endure lies in engineering principles. Concepts like separation of concerns, modularity, and proper architectural patterns remain timeless. AI can assist, but it takes human judgment and experience to apply these fundamentals wisely.
In the end, the best outcomes come from partnership: automation providing speed, and human expertise ensuring software is built to last.
✨ At SGDC Haraya, “change lives” Isn’t just about what we build for others it’s also about how we grow ourselves. For me, returning to development with AI has been a reminder that no matter how long you’ve been in this field, there’s always room to learn, adapt, and rediscover the joy of building.
コメント