Martinkysel.com Website Review

when i was looking for a simple programming task that i can use to evaluate candidates, i had the following criteria in mind: i went through my list of simple problems (i only considered the easy ones) and i stumbled across pairs solution, the specification can be expanded, we can talk about sorting, hashing, hash collisions, big-o notation and cases when the my first interview was planned and i was thrilled. i would use the challenge and it would be great! the first three candidates i exposed to solve it flawlessly. they both accepted the offer and i have been thrilled to work with them ever since. there have been others that got stuck in the weeds but nevertheless showed an ability to think. kudos to those! i will expand on the definition of flawlessly below. i adjust the -pas*- criteria accordingly to the person’s experience (people who graduated a long time ago might not remember what big-o is), title (did they code in the last job?), future job duties (are they supposed to code?) and the prowess with which they claim that they know how to program. support engineers that have had limited exposure to python and shell are fine with something that resembles code. people that have been in the industry for a decade better be ready to go through at least 3 iterations of increasing complexity. find the number of ways that a given integer, x , can be expressed as the sum of the nth powers of unique, natural numbers. for example, if x = 13 and n = 2, we have to find all combinations of unique squares adding up to 13. the only solution is 2^2 + 3^2. this solution does not use dp or memoisation, but it does not time out. we know that it needs to explore all p! possible combinations where p is floor(sqrt(x, n)). since maximum x is 1000, this should still compute in real time. keep in mind that all integers in the solution have to be unique. so for x = 3, the solution 1^2 + 1^2 + 1^2 is not acceptable. did i mention that i dislike both np-hard problems and recursion? no? so, now you know.