My AI Portfolio is LIVE on Vercel… because Google Cloud Run defeated me (this time).
I spent the whole of last week pushing my limits while participating in the Google AI “New Year, New You” Portfolio Challenge. My goal was to transform a simple, static React portfolio into a full-stack application with a smart AI Terminal Assistant.
I wanted to build something that could track real-time location, solve complex math problems using LaTeX, and even change the website’s theme on command.
The challenge required deploying the project to Google Cloud Run.
Here’s the reality, the app was built, the code worked, and the AI was responsive, but I missed the submission deadline.
So what happened?
It wasn’t a coding issue. It was a deployment bottleneck.
My portfolio combines a React frontend with a custom Node.js backend. While this worked seamlessly on localhost, deploying this hybrid architecture to Google Cloud Run was a completely different challenge.
Unlike Vercel, which auto-detects the build process, Cloud Run requires a carefully configured containerized environment. I spent the final 48 hours wrestling with Docker builds, port mappings, and infrastructure issues instead of polishing the product. I simply ran out of runway.
Instead of letting the project die in a folder, I made a decision: get it live, no matter what.
I pivoted to Vercel, and within minutes the project was online and fully functional. It didn’t meet the hackathon’s hosting requirements, but it worked for users, and that mattered more to me.
I may have missed the challenge, but I walked away with a fully functional, live AI portfolio and some hard-earned DevOps lessons.
Link to GitHub Repo: https://github.com/lexycon002/web_portfolio/tree/main
Demo: https://hammad-mu.vercel.app/
Top comments (0)