Building a "Living" Portfolio with Gemini 2.5 Flash and Google Antigravity

Rohan Lopes — Sun, 18 Jan 2026 19:48:38 +0000

This is a submission for the New Year, New You Portfolio Challenge Presented by Google AI

About Me

I am a Senior Integration Architect and AI Specialist with over 15 years of experience in middleware, cloud platforms, and AI-driven development. I specialize in modernizing legacy systems and building advanced, agentic AI solutions.

My portfolio isn't just a static resume—it's a living demonstration of my work. I wanted to express my passion for Actionable AI, moving beyond simple chatbots to tools that actually do things. This site reflects my journey from traditional integration architecture to the cutting edge of Generative AI, showcasing my ability to blend enterprise-grade reliability with innovative AI capabilities.

Portfolio

Live Mirror: rohanlopes.com

Infrastructure: Deployed on Google Cloud Run (labeled dev-tutorial: devnewyear2026) via Firebase App Hosting.

How I Built It

I architected this portfolio as an AI-native web application, moving away from static content in favor of a “living” data-driven experience.

The Development Process

The project was initialized via Firebase and developed using Google Antigravity, an AI-first development environment. Using Antigravity allowed me to rapidly build and test complex AI flows directly within my workspace, creating a seamless transition between writing frontend logic and engineering AI interactions.

The Tech Stack

Framework: Next.js 15 (App Router) and React 19
AI Orchestration: Firebase Genkit utilizing the latest Gemini 2.5 Flash model (v1beta)
Deployment: Firebase App Hosting, leveraging Google Cloud Run for a scalable, serverless backend

Key AI Features

I didn't want AI to just be a bolt-on; I wanted it woven into the fabric of the site:

Dynamic “About Me” Generation

Static bios get stale. My “About Me” section is re-generated on server-side render using a Genkit flow (generateAboutMe). It uses my structured resume data to generate a fresh, 150-word Dynamic Synopsis every time, highlighting different aspects of my experience (from Boomi integration to AI agents).
AI Project Summaries

Each project page uses a Genkit flow (generateProjectSummary) to distill complex technical details into concise, engaging summaries. This helps visitors quickly grasp the Problem, Solution, and my specific Contributions.
“Ask Rohan’s AI Assistant”

I built a RAG-like chatbot (/ask-resume) that allows recruiters and visitors to interview me 24/7. Using Google Genkit, the bot has access to my full resume context to answer specific questions accurately.

What I’m Most Proud Of

I am most proud of the Development Velocity I achieved by combining the Google Antigravity IDE with Genkit.

Rapid Feature Iteration

Even for a streamlined portfolio, the Antigravity IDE was a game-changer. Its AI-native environment allowed me to scaffold the project, build out new pages, and integrate Gemini-powered features in a fraction of the time it would take in a traditional editor. It transformed the way I bridge the gap between prompt design and UI implementation.
Type-Safe AI

By defining input/output schemas for my prompts, I ensured that features like the Project Summaries return predictable, structured data that my UI can render beautifully.
“Living” Content

The Dynamic About Me section feels alive. Seeing a slightly different, high-quality professional pitch generated by Gemini on refresh makes the site feel responsive and personal.
Architectural Reliability

By deploying via Firebase App Hosting to Cloud Run, I maintained an enterprise-grade backend. This setup gives me the reliability of a Senior Architect’s mindset with the speed of a modern AI stack.

DEV Community: Rohan Lopes