Technical Interview Prep

Why active recall beats re-reading notes

The biggest mistake in interview prep is passive review — skimming a list of answers and feeling like you "know" them. Recognising an answer when you see it is not the same as producing it under pressure. Memory research is consistent on this: the act of retrieving an answer from your own head, before checking it, is what builds durable, fast recall. This tool is built around that idea. Every mode forces you to commit to an answer (or admit you don't know it) before the solution appears, which is exactly the muscle a live interview tests.

What's inside the deck

The bank holds 200+ question-and-answer pairs spread across seven categories: Data Structures, Algorithms, System Design, JavaScript, Python, SQL, and Behavioral. These aren't trivia — they're the canonical questions that recur across real screens: hash-table complexity, BFS versus DFS, CAP theorem, closures and the event loop, the GIL, JOIN types, and STAR-method behavioral prompts. You can study the whole deck or filter to a single topic from the dropdown, and a search box matches against both questions and answers so you can drill a narrow subject like "complexity" or "index" in seconds.

The three study modes

The multiple-choice distractors are deliberately drawn from the same category as the correct answer, not random nonsense. That's harder — and more useful — because the wrong options are plausible, forcing you to distinguish "O(log n)" from "O(n)" rather than from something obviously absurd.

How the spaced repetition actually works

This is the feature worth understanding. In Self-Rate mode, marking a card Hard assigns it a priority weight of 2; Easy gives it 1; an unrated card is 0. The deck is then sorted in descending order of weight, so your hardest cards float to the front and you see them first and most often, while cards you've mastered drift to the back. This is a lightweight form of spaced repetition: instead of fixed time intervals, it surfaces the material you're weakest on. The principle is sound — concentrate practice where the gaps are — and because the ranking updates every time you rate a card, the deck continuously re-prioritises around your blind spots.

A worked example

Filter to Algorithms and pick Self-Rate. The first card asks about quicksort's time complexity. You think "O(n log n) average, O(n²) worst," click to reveal, and you were right — mark it Easy. The next card asks about Dijkstra's algorithm and you blank on the fact that it fails with negative edge weights. Mark it Hard. Now that Dijkstra card jumps ahead of the easy quicksort card in the ordering, so on your next pass through the Algorithms deck you'll hit the weak one sooner. Repeat across a few sessions and the deck naturally reshapes itself around what you keep missing.

How to study effectively

1. Do a recognition pass in Multiple Choice first to find which categories feel shaky, then switch to Self-Rate to grind them.
2. Answer out loud or in writing before flipping. Thinking "I roughly know this" is the trap; force a full spoken answer as you would in the room.
3. Be ruthless when rating. If you hesitated at all, mark it Hard. The system only helps if your ratings are honest.
4. Use the category grid at the bottom to track coverage — it shows how many cards in each topic you've seen, so you don't over-study one area and neglect another.

Common mistakes

• Flipping too early. Revealing the answer the instant you read the question turns active recall back into passive reading. Commit first.
• Marking cards Easy to feel productive. Inflated ratings push hard material to the back where you'll avoid it — the opposite of what you want.
• Memorising the answer text verbatim. Interviewers probe; understand why a BST is O(log n) on average and O(n) in the worst case, not just the words.
• Ignoring the Behavioral category. Technical candidates routinely lose offers on STAR-method questions; those cards prep the structure, not just the content.

Privacy and persistence

Your progress and your Easy/Hard ratings are saved in your browser's localStorage, so they persist between visits and the deck remembers what you've struggled with. Nothing is sent to a server — close the tab and come back tomorrow, and your prioritised deck is still waiting.