How andrewekhalel/mlquestions works
Project type and purpose
• This repository is essentially a content‐only, documentation-style project: a collection of Markdown files with interview questions for ML, CV, and NLP engineering roles.
• There is no application code (no “frontend” frameworks, no “backend” services). It’s meant to be consumed as static content—either read directly on GitHub or served through a static-site generator (e.g., GitHub Pages, Jekyll, Hugo, Gatsby).
File structure
• Top level:
• README.md – the master list of 65 Machine Learning & CV questions
• NLP/README.md – the recently added subset dedicated to NLP questions
• There are no “src,” “lib,” or “config” directories, no scripts or tests.
• This suggests a simple content repository rather than a multi-service system.
README for additional insights
• The README describes the purpose (interview prep Q&A) and contribution process (fork, commit questions/answers, pull request).
• No mention of a build process, runtime components, or external integrations beyond GitHub.
a. Content Repository
• Markdown source files (README.md + NLP/README.md) stored in GitHub.
b. CI/CD Build Pipeline (optional but recommended for publishing)
• Trigger: git push / pull request
• Job: Static-site generator (Jekyll, Hugo, Gatsby) converts Markdown → HTML/CSS/JS.
c. Artifact Storage / Hosting
• GitHub Pages, Netlify, Vercel, or similar static-site host + CDN
d. End-user (Developer / Interviewee)
• Browsers (Chrome, Firefox) fetching pages from CDN/host
e. Optional External Services
• Analytics (Google Analytics, Plausible)
• Search (Algolia DocSearch)
• Sponsorship links (e.g., Jobbyo ad banner)
Relationships and interactions
• Optional: Browser sends analytics events to external analytics service.
Architectural patterns & design principles
• “GitOps” for content management: use Git as the single source of truth.
• “Static-site architecture”: no application servers; all dynamic behavior occurs at build or client side.
• Minimal attack surface (security through simplicity).
Relevant technologies/frameworks
• GitHub (version control + hosting)
• Static-site generator (commonly Jekyll or Hugo)
• Continuous Integration: GitHub Actions or similar
• Hosting: GitHub Pages, Netlify, Vercel
• CDN layer (built into hosting)
• Client: modern web browser, responsive design (if a theme is used)
Be detailed but not over-engineered
• Even though there’s only Markdown, break it into discrete pipeline stages.
• How adding a file under NLP/ triggers automatic inclusion in the generated site.
• Contributor workflow separately from reader workflow.