Portfolio Site

A fullstack portfolio and interactive ML/GenAI showcase built with Flask. Features eight self-contained project demos spanning retrieval-augmented generation, computer vision, NLP, unsupervised learning, classification, regression, data visualization, and music data exploration — plus a blog, resume, and contact pages.

Projects

Each project lives in its own GitHub repo, wired in here as a git submodule.

GenAI

Project	Description	Runs On
D&D RAG	Ask natural language questions about D&D 5e — semantic search retrieves relevant passages from the DMG, PHB & MM, then Claude synthesizes a cited answer	Server (sentence-transformers, sqlite-vec, Anthropic Claude)
Text Summarization	Summarize text with 4 algorithms — TextRank, LSA, Luhn, and BART (abstractive)	Server (PyTorch, Transformers)

Machine Learning

Project	Description	Runs On
Digit Recognition	Draw a digit and a neural network identifies it in real-time with confidence scores	Client (TensorFlow.js)
Customer Segmentation	Animated K-means++ clustering with RFM analysis, elbow method, and silhouette scoring	Client (Chart.js)
Mushroom Classification	Compare Random Forest, Gradient Boosting, Logistic Regression, and Decision Tree on the UCI Mushroom dataset	Server (scikit-learn)
Longevity Prediction	Predict age at death from lifestyle habits using four regression models with GridSearchCV tuning and EDA plots	Server (scikit-learn, seaborn)

Data & Visualization

Project	Description	Runs On
Duolingo Visualizer	Dashboard with calendar heatmap and charts for language learning progress	Client (Canvas, Chart.js)
Related Artists	Explore artist collaborator networks via the Discogs API with streaming results	Server (Discogs API, SSE)

Architecture

Projects are self-contained Flask blueprints discovered and registered automatically at startup. Each project lives in its own GitHub repo, mounted as a git submodule under projects/. Each package exposes a PROJECT_META dict and a bp Blueprint — adding a new project requires no manual wiring in the main app.

ml_hub/
├── app.py                     # Flask app factory, portfolio routes, blog registry
├── requirements.txt
├── Dockerfile
├── docker-compose.yml
├── .env.example
├── nginx/
│   └── nginx.conf             # Reverse proxy with SSL
├── static/
│   ├── css/main.css           # Global styles, dark/light theme
│   └── js/                    # Shared JS utilities
├── templates/
│   ├── base.html              # Master layout (topbar, sidebar, theme toggle)
│   ├── landing.html           # Portfolio homepage
│   ├── projects.html          # Project listing grid
│   ├── about.html
│   ├── resume.html
│   ├── contact.html
│   ├── blog.html
│   └── blog/                  # Individual blog posts
└── projects/
    ├── __init__.py             # Auto-discovery and blueprint registration
    ├── customer_segmentation/  # submodule → ZacStryker/customer-segmentation
    ├── digit_recognition/      # submodule → ZacStryker/digit-recognition
    ├── duolingo_visualizer/    # submodule → ZacStryker/duolingo-visualizer
    ├── mushroom_classification/ # submodule → ZacStryker/mushroom-classification
    ├── rag_chatbot/            # submodule → ZacStryker/dnd-rag
    ├── related_artists/        # submodule → ZacStryker/related-artists
    ├── text_summary/           # submodule → ZacStryker/text-summary
    └── work_life_regression/   # submodule → ZacStryker/work-life-regression

Tech Stack

• Backend: Flask, Gunicorn (gevent), Python 3.12
• Frontend: Vanilla JavaScript, Chart.js, TensorFlow.js, Canvas API
• GenAI: Anthropic Claude API (streaming), sentence-transformers, sqlite-vec, RAG pipeline
• ML/NLP: PyTorch, Hugging Face Transformers (BART), NLTK, Sumy, scikit-learn, seaborn
• APIs: Anthropic, Discogs (python3-discogs-client)
• Deployment: Docker, Nginx, Let's Encrypt SSL
• Styling: Custom CSS with dark/light theme via CSS variables and localStorage

Getting Started

Local development

git clone --recurse-submodules https://github.com/ZacStryker/portfolio.git
cd portfolio
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt
python app.py

The app starts at http://localhost:5000.

If you cloned without --recurse-submodules, run git submodule update --init --recursive to pull the project code.

Docker

cp .env.example .env
# Edit .env with your Discogs token
docker compose up --build

Nginx serves on ports 80/443. Flask runs internally on port 8000 behind Gunicorn with a gevent worker and a 300-second timeout for long-running ML inference.

Environment Variables

Variable	Description	Default
ANTHROPIC_API_KEY	Anthropic API key (required for D&D RAG)	—
DISCOGS_USER_AGENT	User-Agent string for Discogs API requests	MLHub/1.0
DISCOGS_USER_TOKEN	Discogs personal access token (required for Related Artists)	—

Routes

Path	Description
/	Landing page
/projects/	Project listing grid
/blog/	Blog index
/blog/<slug>/	Individual blog post
/resume/	Resume
/about/	About
/contact/	Contact
/rag-chatbot/	D&D RAG project
/digit-recognition/	Digit Recognition project
/text-summary/	Text Summarization project
/customer-segmentation/	Customer Segmentation project
/mushroom-classification/	Mushroom Classification project
/work-life-regression/	Longevity Prediction project
/duolingo-visualizer/	Duolingo Visualizer project
/related-artists/	Related Artists project