About this bank

What this is

1,500 multiple-choice questions across four subjects, each written to match the style of a specific exam or interview round, and each carrying a written explanation of why the correct answer is correct. It is a drilling tool, not a textbook — it assumes you have studied and want to find out what you actually retained.

Where the questions come from

Every question is written from scratch against the public syllabus and vendor documentation for its subject, in the style of the exams listed below. Nothing here is copied from a real exam paper — reproducing certification questions violates the exam agreements and, more practically, teaches you the wrong thing. What is modelled is the shape of the questions: the distractor patterns, the scenario framing, the level of detail the examiner expects.

• ☁️ AWS Cloud

  Modelled on: AWS CLF-C02 · AWS SAA-C03 · Cloud Computing university papers · Cloud interview rounds

  100 easy · 100 medium · 100 hard

• 🏅 AWS SAA-C03

  Modelled on: AWS SAA-C03 · Solutions Architect interviews · Cloud architecture reviews

  100 easy · 100 medium · 100 hard

• 🔀 Git & GitHub

  Modelled on: DevOps / SE university papers · Git viva & lab exams · SDE interview VCS rounds · GitHub Actions basics

  100 easy · 100 medium · 100 hard

• ⚙️ MLOps

  Modelled on: MLOps / ML Engineer interviews · AWS MLS-C01 style questions · Applied ML case-study rounds · ML systems design

  100 easy · 100 medium · 100 hard

• 🐍 DSA in Python

  Modelled on: SDE coding interviews · DSA university papers · Competitive programming basics · Python-specific gotchas

  100 easy · 100 medium · 100 hard

What the difficulty levels mean

Easy

One fact, one step. If you have read the material once you should be near 100%. These exist to find the gaps you did not know you had, not to feel good about.

Medium

Two facts combined, or a choice between options that are all technically valid but only one of which fits the stated constraint. This is where most exams sit.

Hard

Scenarios with genuinely plausible distractors, failure-mode reasoning, and trade-offs with no free lunch. In MLOps this level is largely case-study driven; in DSA it leans on complexity analysis and Python-specific semantics.

The AI tutor

Each question already ships with a written explanation. When that is not enough, the tutor button sends that one question, your answer and the existing explanation to Claude and asks for a deeper walkthrough. It is optional, it is per-question, and nothing about your progress or history is sent. If the deployment has no API key configured, the button simply reports that the tutor is unavailable and the site works exactly as before.

Your data

There is no account and no database. Attempts, review scheduling, bookmarks and preferences are stored in this browser's localStorage. That means your progress does not follow you to another device, and clearing site data erases it — so if you study on more than one machine, use the export button on the progress page and import the file on the other one.

Found a wrong answer?

Question banks always have errors, and a confidently wrong explanation is worse than no explanation. Each question has a stable id shown in the review screen — quote it when reporting a problem so it can be found and fixed.