DEV Community

Ignacia Heyer
Ignacia Heyer

Posted on

Ignacia Portfolio engine V!

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

<!-- Tell us a bit about yourself and what you hope to express with your portfolio. --> Based in Gloucester, Massachusetts, I am Ignacia Heyer—a creative technologist, Web3 entrepreneur, and wellness-driven builder.

My work is centered at the intersection of AI, rapid prototyping, brand identity, and decentralized digital experiences. My passion is transforming abstract ideas into clean, scalable systems that are intentional, human-centered, and focused on the future.

This portfolio is not a static website; it is a living, generative system that evolves alongside me. Through it, I aim to convey three key aspects:

  1. Identity: A digital presence that genuinely reflects my voice, energy, and vision.
  2. Clarity: A clear demonstration of who I am, what I build, and the value I deliver.
  3. Momentum: An expression of the speed and creativity I apply to the prototyping process.

Portfolio

<!-- Embed your Cloud Run deployment here using the instructions from https://dev.to/devteam/you-can-now-embed-cloud-run-deployments-directly-in-your-dev-posts-1jk8
Don't forget to include the dev label! -->curl -X POST "https://ignacia-portfolio-engine-v1-473525287303.us-west1.run.app/generate" \
-H "Content-Type: application/json" \
-d '{"task_type": "one_page_pdf"}'

How I Built It

<!-- Share your tech stack, design decisions, and development process. What Google AI tools did you use? --> I restructured my portfolio as an AI-powered engine rather than a conventional website. Instead of manual content creation, I engineered a system capable of generating, on demand and via a single API call:

  • Complete multi-page portfolio websites
  • Single-page, PDF-style resumes
  • Professional pitch decks
  • Judge-ready introductory scripts
  • Social media announcements

This was designed to establish a dynamic identity layer that instantly adapts its format and content for diverse audiences.

Tech Stack Overview

Frontend / Output Layer

- Markdown‑based content generated by Gemini

- Can be rendered into websites, PDFs, or slides

Backend

- Python (Flask) API

- Custom endpoints: /generate, /status, /health, /tasks, /connect, /demo

- Structured logging + error handling

- Versioned service architecture

Infrastructure

- Google Cloud Run (serverless, auto‑scaling)

- Docker container for portability

- Environment variables for secure API key management

AI Layer

- Google AI Studio (Gemini)

- Custom system prompt trained on:

- My bio

- Skills

- Projects

- Brand voice

- Social media links

- Output formats

Technical Architecture Overview

This solution is built upon a modern, serverless technology stack, leveraging Google's AI and cloud infrastructure.

AI and Content Generation

  • AI Engine: Google AI Studio (Gemini) is used as the core AI layer.
  • Customization: The Gemini model is trained via a custom system prompt incorporating:
    • My professional bio and skills.
    • Key projects.
    • My specific brand voice.
    • Social media links.
    • Defined output formats.
  • Output Layer (Frontend): Content is generated in a flexible Markdown format, allowing it to be easily rendered into various final products (websites, PDFs, presentation slides).

Backend and API

  • Framework: A Python (Flask) API manages the core application logic.
  • Endpoints: Essential custom endpoints include:
    • /generate (for content creation)
    • /status and /health (for monitoring)
    • /tasks
    • /connect
    • /demo
  • Development Practices: The backend features structured logging, robust error handling, and a versioned service architecture.

Infrastructure and Deployment

  • Platform: Google Cloud Run provides a serverless and auto-scaling environment.
  • Portability: The application is containerized using Docker.
  • Security: Environment variables are utilized for the secure management of API keys.AI-Powered Portfolio Engine: Design and DevelopmentCore Design Principles

This project was built on five key design decisions, ensuring a flexible, predictable, and professional system:
Generative over Static: The portfolio is designed to evolve automatically, eliminating the need for manual updates and ensuring continuous relevance.
Universal Markdown Output: Markdown was selected as the sole output format for its flexibility and easy conversion into multiple assets:
Websites
PDFs
Slide decks
Social media posts
Clear Task-Based Architecture: Predictability and extensibility are achieved by generating each asset based on a specified task_type.
Human-Centered Tone: A dedicated system prompt enforces a consistent, professional, and engaging voice for every output, ensuring the tone is:
Confident
Clear
Professional
Creative
Future-focused
Serverless Deployment (Cloud Run): This choice provides a robust, zero-maintenance infrastructure, offering:
Instant scaling
A public HTTPS endpoint
Seamless integration with AI Studio
-----Development Workflow

The creation of the AI-powered portfolio engine followed a systematic six-step process:
Identity Definition (System Prompt): A detailed system prompt was drafted to capture my professional identity, including my story, skills, projects, desired tone, and social links.
AI Logic Building (Google AI Studio): The core AI logic was built and rigorously tested in Google AI Studio, iterating on outputs like multi-page websites, pitch decks, PDFs, intro scripts, and social posts until consistency and quality were achieved.
Backend Creation (Flask API): A Flask API was developed to serve as the intermediary, receiving a task, sending it to the Gemini API, and returning the clean Markdown output.
Professional Endpoint Implementation: Standard, production-ready endpoints were added for system management and information: /status, /health, /version, /tasks, /connect, and /demo.
Deployment (Cloud Run): The application was containerized with Docker and securely deployed to Google Cloud Run, utilizing environment variables for security.
Final Testing and Refinement: All endpoints were validated using curl, and the generative outputs were polished to ensure perfection.
-----Google AI Tools Utilized

The engine relies on a powerful combination of Google AI and Cloud technologies:
Tool
Function
Google AI Studio (Gemini)
Used for defining the custom system prompt, task-based generation, multi-format content testing, and overall iteration environment.
Gemini API
The core generative engine integrated into the Cloud Run backend, responsible for creating all portfolio assets, handling structured prompts, and producing Markdown.
Google Cloud Run
The hosting platform for the API, providing serverless scaling, automatic HTTPS management, logging, and deployment infrastructure.

Together, these tools form a modern, flexible, and fully automated AI-powered portfolio engine tailored to my unique identity.

What I'm Most Proud Of

<!-- Highlight the features, design elements, or technical achievements you're most excited about. --> My greatest achievement is the development of a portfolio system that goes beyond simply showcasing my work; it embodies my forward-thinking philosophy. Rather than a traditional, static website, I engineered a dynamic, generative identity layer that constantly adapts alongside my evolution. This project reflects my core belief that design must be adaptive, intentional, and profoundly human. It's an intersection of artistic expression and technical execution—precisely where I find myself doing my most impactful work.

Top comments (0)