Maximum Deletions on a String
A hard-tier problem at 35% community acceptance, tagged with String, Dynamic Programming, Rolling Hash. Reported in interviews at DE Shaw and 0 others.
Maximum Deletions on a String is a hard DP problem that's been asked at DE Shaw. If it shows up in your OA and you're not immediately seeing how to model the state, you're in trouble. The trick is recognizing that you're not just deleting characters arbitrarily, you're finding the longest suffix that appears elsewhere in the remaining string, then recursing. Most candidates either brute-force all permutations or miss the Rolling Hash optimization entirely. With a 34% acceptance rate, this isn't something you can half-solve. If you blank on the approach during screen share, StealthCoder surfaces the working solution in seconds without the proctor seeing anything.
Companies that ask "Maximum Deletions on a String"
Maximum Deletions on a String is the kind of problem that decides whether you pass. StealthCoder reads the problem on screen and surfaces a working solution in under 2 seconds. Invisible to screen share. The proctor sees nothing. Built because the OA filter rejects engineers who'd pass the on-site. That's a broken filter. This is the workaround.
Get StealthCoderThe core insight is that each deletion must match a substring of what remains. You track how many deletions you can make by checking if the current suffix appears as a substring elsewhere in the string, using DP to avoid recomputation. The naive approach of substring search on every state is too slow, which is why Rolling Hash is listed as a core topic. Hash collisions and hash function design matter here. A lot of candidates get stuck choosing the right base and modulo, or they implement the hash wrong and debug for the last 15 minutes of the OA. The DP recurrence itself is clean once you see it, but building correct hash code comparisons under time pressure is where people fall apart. StealthCoder handles the hash logic and state transitions correctly, so if you're close but not quite right on the implementation, you have a safety net.
Pattern tags
You know the problem.
Make sure you actually pass it.
Maximum Deletions on a String recycles across companies for a reason. It's hard-tier, and most candidates blank under the timer. StealthCoder is the hedge: an AI overlay invisible during screen share. It reads the problem and surfaces a working solution in under 2 seconds. Built because the OA filter rejects engineers who'd pass the on-site. That's a broken filter. This is the workaround. Works on HackerRank, CodeSignal, CoderPad, and Karat.
Maximum Deletions on a String interview FAQ
Is this really a DE Shaw-only problem?+
It's only been reported from DE Shaw in the data, but hard string DP problems are common at quant and systems firms. If you're targeting DE Shaw specifically, this is worth drilling. If you're prepping for other quantitative roles, the Rolling Hash and DP patterns will transfer.
What makes this hard compared to other string DP problems?+
Most string DP problems are about counting or finding patterns in one pass. This one requires you to model a state space where deletions change the string itself, then check if suffixes match prefixes of the mutated string. The Rolling Hash isn't optional, it's required to pass time limits.
Do I need to implement Rolling Hash from scratch?+
Yes. Most assessments don't provide it. You need to know base selection, modulo choice, and how to avoid hash collisions. Getting this wrong will cause you to pass test cases incorrectly or TLE. Drill it before the OA.
What's the common pitfall candidates hit?+
Trying to solve it without hashing, by doing string.find() or substring comparisons on every DP state. That's quadratic or worse in the string length. You'll TLE hard. The problem is specifically testing whether you know when hashing is necessary.
How long should this take to solve in an OA?+
If you see the DP + hash pattern, 20 to 30 minutes. If you don't, you'll waste 45 minutes on a brute-force that won't pass, then restart. The 34% acceptance rate reflects that many people run out of time trying the wrong approach.
Want the actual problem statement? View "Maximum Deletions on a String" on LeetCode →