DEV Community

Rachel Wang
Rachel Wang

Posted on

Rachel Wang's Portfolio - Built with Gemini & Cloud Run

#devchallenge #googleaichallenge #portfolio #gemini

New Year, New You Portfolio Challenge Submission

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

About Me

I'm Wei-Hsien (Rachel) Wang, a recent MS in Business Analytics graduate from UC San Diego (Class of 2025). I specialize in building production-grade AI/ML systems, with hands-on experience in LLM applications, GraphRAG architectures, and scalable data pipelines. My career has taken me from research labs to enterprise AI deployments, and I wanted my portfolio to reflect that journey—showcasing not just what I've built, but how I think about solving real-world problems with AI.

Portfolio

Live URL: https://portfolio-site-554933605441.us-central1.run.app

How I Built It

Tech Stack:

  • Frontend: HTML5, CSS3, Vanilla JavaScript with smooth animations and responsive design
  • Backend: FastAPI (Python) serving both static files and a chat API
  • AI Integration: Google Gemini 3 Flash Preview for intelligent portfolio Q&A
  • Hosting: Google Cloud Run with automated CI/CD via Cloud Build
  • Infrastructure: Docker containerization, Artifact Registry for image storage

Design Decisions:

  1. AI-Powered Chat Assistant: I integrated Gemini 3 Flash Preview to create an interactive chat experience where visitors can ask questions about my background, projects, and skills. The model is fed the entire portfolio HTML as context, allowing it to provide accurate, contextual responses based strictly on my actual experience.

  2. Clean, Professional Design: I chose a minimalist design with Inter font, smooth CSS animations, and a timeline-based experience section. The goal was to make the content easily scannable while maintaining visual interest through subtle interactions.

  3. Performance-First Architecture: Using FastAPI for the backend allows for fast response times, while Cloud Run's auto-scaling ensures the site can handle traffic spikes. The static files are served efficiently, and the chat API is only called when needed.

  4. Responsive Layout: The portfolio adapts seamlessly across devices using CSS Grid and Flexbox, ensuring a great experience whether you're viewing on mobile, tablet, or desktop.

Development Process:

  • Started with content strategy—organizing my experience into clear categories (AI Agents, Modeling, Data Engineering)
  • Built the frontend with semantic HTML and progressive enhancement
  • Integrated Gemini API with a system prompt that understands my portfolio context
  • Containerized the application with Docker for consistent deployment
  • Set up Cloud Build for automated deployments on every git push
  • Added the required dev-tutorial=devnewyear2026 label for the hackathon

What I'm Most Proud Of

1. The Gemini-Powered Chat Experience: Rather than just listing my accomplishments, I wanted visitors to be able to have a conversation about my work. The chat assistant can answer questions like "What AI projects has Rachel worked on?" or "Tell me about her experience with GraphRAG" by intelligently parsing the portfolio content. It's like having a 24/7 recruiter who knows my background inside and out.

2. Production-Ready Architecture: This isn't just a static site—it's a full-stack application with proper error handling, CORS configuration, environment variable management, and automated CI/CD. It demonstrates the same engineering practices I use in my professional work.

3. Storytelling Through Structure: I organized my projects into meaningful categories (AI Agents, Modeling & Experiments, Data Engineering) that tell a story about my technical evolution. Each project includes not just what I built, but the impact it had—like reducing task discovery time by 65% or improving recommendation like-rates by 20%.

4. Technical Depth: The portfolio showcases real production systems I've built—from processing 500K+ enterprise emails with LangGraph to deploying RAG systems on Azure. These aren't toy projects; they're systems that solved actual business problems.

5. Seamless Google Cloud Integration: The entire deployment pipeline leverages Google Cloud's ecosystem—Cloud Run for hosting, Cloud Build for CI/CD, and Artifact Registry for container management. It's a testament to how powerful Google's AI and cloud tools are when used together.

Connect with me:

Top comments (0)