ML System Design

Welcome folks 🎉 — this is your single-pane-of-glass for everything ML System Design. The content below is organized to help you learn fast, revise faster, and answer interview questions confidently. Use the left sidebar to browse full notes; use the quick links here to jump straight into the most-used sections.

🧭 Start Here: Introduction
Problem framing, funnels, evaluation strategy.
 ⚡ Performance & Capacity
SLA, latency, scaling & funnel tradeoffs.
 📥 Training Data
Collection modes, bias, bootstrapping & augmentation.
 🧪 Online Experimentation
A/B testing, power, backtesting, long-runs.
 🔗 Embeddings
Word/image/user/item vectors & two-tower models.
 🔁 Transfer Learning
Feature-extract vs fine-tune strategies.
 🔍 Model Debugging
Offline vs online mismatch, drift, PSI, KS tests.

✨ Why this landing page exists

  1. Reduce friction — everything you need to review or teach is one click away.
  2. Study friendly — flashcards, short-theories, deep Q&A, formulas and cheat-sheets are present inside each page.
  3. Interview ready — each topic includes Q&A blocks formatted as collapsible details for quick mocks.

🔎 Quick Navigation

  • Use the cards above for direct jumps.
  • Use the sidebar to see the full table of contents and progress across parts.
  • Use your browser’s Find (Ctrl/Cmd+F) inside any page to jump to formulas or interview Q&A quickly.

🧰 Study Workflows (Pick one)

Fast Revision (10–20 min)
  1. Open Introduction → scan flashcards.
  2. Open Performance & Capacity → read the funnel & SLA bullets.
  3. Finish with Cheatsheet at bottom of any page for last-minute recall.
Deep-Dive (60–90 min)
  1. Read Short Theories for the topic.
  2. Expand 3–4 Q&A blocks and write short notes.
  3. Attempt a mock interview question using the Interview Q&A prompts.
Interview Simulation
  1. Pick a top-level problem from Introduction.
  2. Use Performance & Capacity + Training Data pages to design answer; rehearse aloud.
  3. Time yourself and refine using the Model Debugging checklist.

📈 Progress & Tips

  • Tip: After reading a page, mark it Done by adding a tiny note in your personal reading tracker.
  • Tip: For reproducible features, check Model Debugging → Feature-logging section.
  • Tip: Use the Online Experimentation notes when planning A/B experiments — contains power/sample tables.

🧾 Printable Cheatsheet

If you want a printable revision sheet, open any topic and use Print → Save as PDF. For best results, print the Cheatsheet sections only (they are near the bottom of each topic page).

🙋‍♀️ Contribute / Edit

Want to update a note, add an example, or include a new page? Drop the new Markdown in /content/snapshots/ml-system-design/ and use the same cards shortcodes — this landing page will auto-link if the url matches.

❤️ Quick CTA

Start now: Click the “🧭 Start Here: Introduction” card above — 3 minutes there will give you a map of the whole course.

Made with ❤️ for engineers who like tidy systems and fast recall.