An immersive, neural-styled portfolio powered by Google Gemini 2.5 Flash. Built 100% using Google AI Studio & Antigravity, it acts as a gateway to my ecosystem of deployed AI applications, featuring a context-aware "AI Ambassador" that guides visitors through the nodes of my experience.
I am Giuseppe, a 46-year-old entrepreneur from Rome. For 20 years, I ran physical businesses, but in 2023, managing a major web platform (Mercato Trionfale) reignited my passion for Python. I had the vision, but traditional coding was too slow.
When Google AI Studio was released, everything changed. It allowed me to bridge the gap between business logic and complex Full-Stack engineering. This portfolio is not just a showcase; it is the proof of this transformation. It visualizes my career not as a list, but as a Neural Network, where every "node" is a real, functioning application I built—from Music EdTech to Retail Management—demonstrating how RAG, OCR, and multimodal AI are evolving software as we know it.
How I Built It (The "AI-First" Methodology)
This project is a manifesto for the new era of AI-Assisted Development. I did not write this code alone; I acted as the Architect orchestrating powerful AI agents.
The "Dual-AI" Workflow:
Google AI Studio (The Architect): I used AI Studio to generate the core architecture of the Backend (FastAPI/Python) and the Frontend (React/Vite). It handled 80% of the logic, data structures, and API integrations.
Antigravity (The Specialist): As the project grew in complexity, I used Antigravity for two specific tasks:
Deep Debugging: Solving complex state-management bugs that appeared when the application scale increased.
Advanced Styling: Creating the intricate "Neural Storm" CSS animations, neon glows, and the "Cerebral Cortex Access" gate that you see in the UI.
The Tech Stack:
Brain: Google Gemini 2.5 Flash (via Dynamic Context Injection).
Core: Python, FastAPI, React, Framer Motion.
Security: A "Bring Your Own Key" system stored locally in the browser, ensuring user privacy.
The Ecosystem (What the Neural Nodes Connect To)
The portfolio is the central hub (The Synapse) that connects to my live production apps:
Mercato Trionfale: A management ecosystem for 275 physical stores (Active for 3 years).
BeyerOp101 & PianoHiTech: AI-driven Music Education platforms using Audio processing and Generative AI to accelerate learning.
Synapsis Lab: The "Motherbase" where we test and launch new vertical solutions.
Challenges I ran into
- The "Black Box" of Complexity: Building with AI is fast, but when a project grows, maintaining control over thousands of lines of code is hard. The challenge was keeping the "Human in the Loop." I had to learn how to guide the AI to fix its own regressions using Antigravity, effectively becoming a "Prompt Architect" rather than just a coder.
- Privacy & Security: I wanted a public demo that didn't bankrupt me. I implemented a client-side "Cerebral Cortex Access" gate. The user inputs their own Google Gemini API Key, which travels encrypted to the temporary backend memory and is never stored on my server. This creates a secure, zero-cost architecture for me.
ps.Many of our applications adopt the same technology of inserting the personal key in the front side and because we think that in the next few years each user will have his own personal key, we are testing the implementation of this system thanks to Google.
Achievements I'm Proud of
Zero-Code-Handwritten: Demonstrating that a complete, animated, and complex app can be built entirely through prompts and AI agent orchestration.
The "Living" Experience: The transition from a static CV to a "Neural Journey" where the AI acts like me, answering questions about my life and my code in real time.
The CSS Visuals: The visual impact of the "Neural Storm" and the "Gate" (created with Antigravity) is an example of the agent's high-level CSS work that truly amazed us.
What's next
I believe OCR, TTS (Text-to-Speech), and STT (Speech-to-Text) are the future of interaction. I am currently integrating a "Voice Mode" so users can literally speak to my digital brain instead of typing, making the experience fully multimodal.
Built With
(Google-AI-Studio),(Gemini-2.5-Flash),(Antigravity) (FastAPI) (React Python) (Generative-UI)
https://github.com/thc791/portfolio-2026-ai-metro
Top comments (0)