For many years, I was a “hands-on keyboard” kind of software engineer. I’ve written millions of lines of code — in languages that are both still loved (hello, C#) and those we pretend never existed (looking at you, VB6). I’ve spent late nights debugging mysterious 500 errors, duct-taping microservices together, and chasing race conditions like a developer-themed episode of CSI.
But then… management happened.
Gradually, I shifted away from writing code to writing strategy docs. I traded commit messages for slide decks. And while I genuinely enjoy leading teams and helping people grow, I started missing the code. But more than that — I started noticing how everything in our world was changing.
Fast-forward to now, and AI has absolutely exploded onto the scene, especially in software development. We’ve seen the rise of code-completion tools, chat-based assistance, and more recently, Agentic AI — AI that doesn’t just answer your questions, but acts, plans, and builds software like a junior developer (hopefully one that doesn’t break the build every Friday afternoon).
👀 So… Can AI Actually Code Better Than Me?
Some folks believe AI will be the glorious end of all software engineers. Others say the quality is awful and you spend more time wrestling the AI than coding. The truth? It’s somewhere in the middle, like pineapple on pizza (don’t @ me).
But that got me thinking:
Can I write better code than AI?
More importantly: Can I teach AI to code like me?
What if I treated AI like a junior developer — someone who needs onboarding, mentorship, a few nudges in the right direction, and a decent naming convention? What if I could turn it into a digital version of myself… but faster, more consistent, and with better coffee tolerance?
That’s the mission I’ve set out on.
🛠️ The Mission: Build Software with AI. Like, All of It.
In this blog series, you’ll follow my slightly unhinged but deeply curious journey into Agentic AI. I’ll attempt to build an actual SaaS product using AI at every step — not just code generation, but all the disciplines that go into building real software:
🧩 Disciplines I’ll let AI help with:
- Requirements gathering and documentation
- Architecture and technology selection
- UI/UX design (sorry Figma, you’re not safe)
- Frontend and backend code
- Infrastructure-as-code (IaC)
- CI/CD pipelines
- Testing strategies
- Monitoring & observability
- DevOps practices
- Security considerations
- Product marketing and branding
- …and even writing blog posts like this one 🙃
Yes, I’ll let AI write about itself. It’s basically very advanced bragging at this point.
🧪 So, What Am I Building?
Good question. I thought about building some wildly disruptive AI startup to take over the world — or at least get a few investors to email me back. But in the end, I chose something a bit more selfish:
I’m building a tool that I genuinely want to use.
You’ve probably heard the term Platform Engineering being thrown around lately — and a lot of companies are now setting up internal developer portals to centralize docs, tooling, environments, self-service deployments, and so on.
There are a few solutions already out there like Backstage (from Spotify) and Port. They’re good, but none of them made me want to hug my laptop. So I’m going to build my own developer portal.
… From scratch.
… As a SaaS.
… With as much help from AI as I can possibly get.
💡 Introducing DevBridge.ai — My AI-Powered Developer Portal
Here’s the elevator pitch my good friend ChatGPT and I came up with:
The Smart Developer Portal for Modern Engineering Teams
DevBridge.ai is an intelligent, self-updating developer portal that brings clarity and automation to modern software development.
It connects to your existing tools, automates documentation and ownership, and gives teams a unified view of everything they’re building and running.
And here’s what it will (hopefully) do:
Auto-generated service catalog
Automatically discovers and categorizes all services, apps, scripts, and jobs.
Ownership detection
Identifies responsible teams or individuals using patterns, commits, and metadata.
AI-generated documentation
Creates and maintains clear, structured documentation from your codebase.
Semantic search & AI assistance
Search everything using natural language and get contextual explanations instantly.
Readiness checks & policy enforcement
Automatically validates services against team-defined standards and best practices.
Dashboards & overviews
Visualize service health, team activity, and development velocity in real time.
Collaboration notifications
Sends proactive updates about changes, risks, or missing data to Slack or Teams.
🧭 What’s Next?
In the next part of this blog series, I’ll be choosing the tech stack, setting up the project, and letting AI take its first baby steps toward becoming the intern of my dreams (or nightmares).
Can AI learn to code like me?
Can I keep my sanity through it all?
Let’s find out together.
👉 Next up: Tech stack, architecture, and my first AI commit.
Want to follow the full journey?
Follow me on LinkedIn, bookmark this page, or just yell “DevBridge!” into the wind — I’ll hear it. 💨
Leave a Reply