CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
Project Overview
This is a personal website built with Quarto, a publishing system for scientific and technical content. The site features blog posts about machine learning, data science, programming, and technical tutorials, along with paper summaries, projects, and personal notes.
Technology Stack
- Quarto: Static site generator for scientific and technical publishing
- Python: For notebooks, scripts, and data analysis examples
- Jupyter Notebooks: Interactive content for technical posts
- HTML/CSS/JavaScript: Generated output for web deployment
- GitHub Pages: Hosting platform (deploys to
docs/
directory)
Development Commands
Building the Site
quarto render
This renders all .qmd
and .ipynb
files to HTML in the docs/
directory.
Preview During Development
quarto preview
Starts a local development server with live reload.
Publishing
The site auto-deploys to GitHub Pages from the docs/
directory when changes are pushed to the main branch.
Project Structure
Content Organization
- Posts (
posts/
): Technical blog posts organized by topic (e.g.,nlp/
,python/
,airflow/
) - Papers Summaries (
papers-summaries/
): Research paper summaries and notes - TIL (
til/
): “Today I Learned” short notes organized by technology - Projects (
projects/
): Portfolio project descriptions - Books/Misc Notes (
books-summaries/
,misc-notes/
): Additional content sections
File Types
.qmd
: Quarto markdown files (main content format).ipynb
: Jupyter notebooks (for code-heavy technical content).md
: Standard markdown (some legacy content)_metadata.yml
: Directory-specific metadata and configuration
Key Configuration
_quarto.yml
: Main site configuration, navigation, themes, and build settingsstyles.css
: Custom CSS styling_citation.qmd
: Citation format included in all pages
Content Creation Workflow
- Blog Posts: Create in appropriate
posts/[category]/
subdirectory - Notebooks: Use for interactive, code-heavy content with visualizations
- Images: Store in topic-specific
images/
subdirectories or mainimages/
folder - Metadata: Include proper YAML frontmatter with title, date, categories, and features
Development Notes
- The site uses Quarto’s listing feature for automatic post organization
- Search functionality powered by Algolia
- Comments system integrated via Giscus
- Google Analytics tracking configured
- Dark/light theme support via Bootstrap themes
- Social media integration (Twitter, LinkedIn, GitHub)
Python Dependencies
The site includes Python scripts and utilities primarily for: - Bandit algorithm implementations and plotting (posts/scripts/
) - BPE tokenizer implementation (bpe.py
) - Data analysis examples in notebooks - Machine learning tutorials and implementations
Most Python code is self-contained within individual post directories or serves as educational examples rather than production dependencies.