Spaced repetition for coding interviewsBeta

Stop forgetting
problems you've
already solved.

Explain it, code it, defend the Big-O. An AI coach grades every LeetCode session and schedules the next review for the exact moment you're about to forget.

Start practicing Read the blog

recur.app / today / two-sum

⌘↵Submit

✓ Explain✓ Code3 · Big-O

Problem

1. Two Sum

Time is O(n) since it's one pass and each lookup is amortized O(1). Space is O(n) for the hash map in the worst case, when no pair resolves early.▍

Coach

Grading

You nailed the single-pass insight. What's the worst case if every key collides in the hash map?

Then lookups degrade to O(n), so the whole thing becomes O(n²).

Exactly. That covers it, grading the session now.

Coach verdict Excellent · next review in 6d

ExplanationClear, spotted the single-pass insight

CodeClean, idiomatic, first try

Big-ONailed the collision worst case

Why spaced repetition

Forgetting is predictable.

A century of memory research maps the same curve: without review, most of what you learn is gone within days. Recur tracks that curve per problem and surfaces the next review right before recall slips.

The result: fewer problems, deeper retention, and the confidence to recall a pattern under pressure, not just recognize it.

~66%

Forgotten within a day without review

Reviews to push a problem past a month

90%

Recall the scheduler is tuned for

Ebbinghaus (1885), replicated by Murre & Dros (2015) · intervals from SM-2 defaults

Recall over time

no review with recur

How it works

One problem, three layers of understanding.

You don't really know a problem until you can explain it cold. Recur walks you through three checkpoints, then the coach grades the whole session.

Explain in plain words

No code. No symbols. Describe the approach like you're whiteboarding to a teammate. The coach catches vague hand-waves.

Your explanation

Two pointers from each end…vague… and move whichever is smaller.▍

Write the code cold

Paste your solution or type it from scratch. The coach reads it like a reviewer would and flags what's off against the canonical pattern.

def two_sum(nums, target):

seen = {}

for i, nin enumerate(nums):

✓ correct

Defend the Big-O

State time + space. Then the coach probes edge cases like collisions, recursion stack depth, and sort ordering until your reasoning holds up.

TimeO(n log n)

SpaceO(1)

Step 04The part that actually sticks

The coach grades the session.

Recur's coach asks the questions a senior interviewer would: trade-offs, alternatives, what breaks at scale. Then it grades the session (Fail, Shaky, Good, Excellent) and that grade decides exactly when the problem lands back on your queue.

You said "two pointers" — but didn't say why sorting first is required. Walk me through it.

Without sorting we lose monotonicity, so shrinking either side wouldn't be safe.

Exactly. That's the reasoning an interviewer wants to hear. Wrapping up the session.

Pricing

One plan. Everything included.

Try everything free for 14 days. No card required, cancel with one click.

14-day free trial

Recur Pro

$14/ month

Start 14-day free trial

Unlimited practice sessions

Reviews timed to land right before you forget

An AI coach that probes you like an interviewer

Progress tracking & analytics (coming soon)

Interviews are
recall tests.

Solve fewer problems. Remember every one. Walk into your next interview without the cold sweat.