This is a submission for the New Year, New You Portfolio Challenge Presented by Google AI
About Me
I am a Computer Science graduate from Adamson University with over 8 years of technical experience across the IT spectrum—spanning DevOps, QA, and my current role as a Business Insights Analyst at TaskUs. At 31, I’ve realized my true passion: the intersection of AI and Automation.
My portfolio, Cyber-Cottage, represents the collision of two worlds: Cyberpunk and Cottagecore. It expresses my drive for a high-tech career while staying rooted in the simple joys of nature and crafting. With my birthday on February 8th, I am choosing this moment to officially pivot into the field of AI & Automation, using intelligent systems to bridge the gap between complex data and real-world efficiency.
Portfolio
Interact with Aura, my AI Greenhouse guide. My skills and projects "bloom" as you engage with the interface! Aura is powered by Gemini 1.5 Flash and is trained on my professional history and project documentation.
How I Built It
Building this was a hands-on exercise in end-to-end AI systems architecture.
The Tech Stack
- Backend: Node.js with Express to securely handle API requests.
- Frontend: Tailwind CSS (Aesthetic) and Alpine.js (Reactive State).
- AI Core: Gemini 1.5 Flash via the Google Generative AI SDK.
- AI-First Dev: Developed within Antigravity, Google’s AI-first IDE, allowing for rapid iteration and debugging.
Architecture & Security
To keep my Gemini API key safe, I architected a server-side proxy. By using a Node.js backend to communicate with Google’s servers, sensitive credentials never reach the client’s browser.
The Brains: Contextual System Prompting
- I engineered Aura's knowledge through a structured system prompt:
- Identity Injection: A witty "AI Greenhouse Guide" persona.
- Knowledge Base Partitioning: Structured data regarding my history at TaskUs, Amdocs, and Accenture.
- Zero-Cache Architecture: Configured headers to ensure every interaction is fresh and reactive.
Automated Deployment
The app is containerized with Docker and deployed on Google Cloud Run.
# Final Deployment Command (Sanitized)
gcloud run deploy ai-portfolio \
--source . \
--region asia-southeast1 \
--set-env-vars EMAIL_USER="<GMAIL>",EMAIL_PASS="<APP_PASSWORD>",GEMINI_API_KEY="<KEY>" \
--labels dev-tutorial=devnewyear2026 \
--allow-unauthenticated
What I'm Most Proud Of
"Every 'Signal Lost' eventually became a 'Transmission Successful.'"
I am most proud of my learning progress and resilience. During this challenge, I faced numerous syntax errors and complex deployment hurdles—from Docker registry mismatches to GCP IAM permission loops. In the past, these technical roadblocks might have been discouraging, but this time, I leaned into the process.
I spent nights retrying, debugging, and learning alongside my AI agent—treating every error not as a failure, but as a "level-up" in my technical understanding.
Key Technical Wins:
Resilient Debugging: Resolved critical CI/CD pathing issues to ensure a seamless "Push-to-Deploy" workflow.
Agentic Orchestration: Successfully moved from "chatting" with AI to building with it, orchestrating agents to deliver a production-ready cloud system.
A Personal Milestone: Successfully launching this portfolio and GreetStyle AI (my generative AI card app) is the best early birthday gift I could give myself.
I’m entering this new year of life with a re-ignited passion for building.
I'd love to know what you think! Feel free to comment on your suggestions and observations!
Thanks for viewing!
Top comments (0)