A founder brought me in to fix a few bugs in an MVP built with Lovable. What looked like a small ...A founder brought me in to fix a few bugs in an MVP built with Lovable. What looked like a small ...
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
A founder brought me in to fix a few bugs in an MVP built with Lovable.
What looked like a small cleanup quickly turned into an architecture problem.
I found: • Unused backend services • Duplicate state management • Real-time chat built with long polling
Instead of patching symptoms, I rebuilt the foundations: simplified state management, redesigned the Supabase integration, removed unnecessary abstractions, and cleaned up the architecture.
The result?
Around 60% less code with the same functionality.
And I still used AI throughout the process.
AI wasn’t the problem. Engineering judgment was the difference — knowing what to keep, what to remove, and what needed to be redesigned.
Tools like Lovable, Cursor, and Replit make MVPs easier than ever to build.
But production software still needs solid engineering.
When an MVP becomes harder to change than it was to build, another prompt usually isn’t the answer.
Post image
Back to feed
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started