🧠 I Solved 700+ LeetCode Problems And Still Blanked in My OA. So I Built the Fix (Free Tool Inside)
Be honest has this happened to you?
You solve "Number of Islands" cleanly. Two months later, you're staring at the exact same pattern in an actual OA or Interview, and your brain just... doesn't fire. Not because you're bad at DSA. Because solving a problem once was never going to make it stick.
This is the part nobody talks about when they tell you to "just grind 500 problems." Grinding gets you exposure. It doesn't get you retention. And retention is the only thing that matters when you're sitting in front of a timer for Amazon, Atlassian, Uber, or Google's OA.
The actual science behind it (no fluff)
Human memory decays on a predictable curve. The fix isn't "revise more" it's "revise at the right intervals." This is the same principle behind Anki and Duolingo: the SM-2 spaced repetition algorithm. The logic is simple:
- Recall a problem easily → the system pushes your next review further out.
- Struggle or fail to recall → the interval resets and tightens.
That's it. Your revision schedule adapts to your memory, not a generic "redo this sheet every week" plan that wastes time on stuff you already know and barely touches the stuff you're about to forget.
So I built AlgoRetention
It's a full-stack tool that applies exactly this to DSA prep:
- Auto-scheduled re-solves - every problem you log gets a "next review" date based on how well you actually recalled it, not a fixed calendar.
- Failure-first logging - problems you struggled on get prioritized back into your queue faster than ones you nailed.
- Pattern tracking - so you're not just revisiting random problems, you're reinforcing the pattern that's weak (sliding window, DP on subsets, whatever it is).
- A dashboard that shows you what's about to decay - instead of guessing what to revise the night before an OA.
I built this because I was prepping for product-company SDE roles and kept hitting the same wall: comfortable in practice, shaky under real pressure, because the problems I'd "solved" weeks ago weren't actually retained they were just recently seen.
Why this matters specifically for OAs
Company OAs (Amazon, Atlassian, Uber, Microsoft, Google — basically anyone running timed pattern-recognition tests) don't reward "I solved this once in March." They reward instant pattern recall under a clock. A spaced-repetition system means the problems you're most likely to forget are exactly the ones surfacing in your queue the week before your test — not buried under 1000 other solved problems you'll never revisit.
Try it
It's free: https://algo-retention.vercel.app/
I'm still actively building this (currently in active placement prep myself, so this is being battle-tested daily, not built and abandoned). If you try it, I'd genuinely value brutal feedback what's missing, what's clunky, what would make you actually use it daily instead of bookmarking it and forgetting it exists.
Good luck with your prep. See you in the OA review threads. 🚀Bold