My smoother interviews began with a technical phone screen, then a comfortable, practical DSA-style round. It included a regression-model critique and a subset-sum-style problem. I had recently worked through similar problems, so those patterns were already familiar.
I then had a codebase-style evaluation around PDF evaluation logic. The structure felt familiar because I had previously reviewed that sort of architectural breakdown. The questions seemed to test whether I could reason about an existing system rather than build everything from nothing, and the process felt efficient and oriented toward learning.
In a different set of similar ML engineering interviews, the pacing felt very different. A project discussion was rushed, followed by a time-boxed and oddly disconnected question about calculating the difference between two dates. That was frustrating because it did not feel tied to algorithmic ML engineering depth. I declined an offer from the smoother journey, and I was left thinking that fit and tone can matter as much as raw technical ability.
Location: United States. Overall feedback: positive. Offer status: got offer.
Discussion
Loading comments…