SDE intern(6 month)
Interview Date
14-08-2026
Result
Rejected
Difficulty
Medium
Rounds
02
Drive Type
Off-Campus
Topics asked
Detailed experience
I applied for Amazon's six-month SDE internship through the off-campus route. The first stage was the online assessment, and it had a mix of debugging, multiple-choice questions, and coding. The coding section had two problems. One was based on comparator-style sorting, while the other involved finding the longest substring with exactly K distinct characters. I was able to pass all the test cases, so I moved ahead to the interviews. This structure is also reported in an India-based candidate experience for the same six-month internship. My first technical interview was almost completely focused on DSA. The interviewer did not spend much time on introductions and moved directly into problem solving. The first problem was based on finding the Nth Magical Number. I started with a brute-force idea, then explained why it would not work efficiently and moved toward binary search. The next problem was finding the minimum XOR among every possible pair in an array. I first discussed the brute-force solution and then improved it by sorting the elements. From there, I also discussed the trie-based approach. The interviewer kept pushing me toward better complexity, so I had to explain not only the code but also why each optimization worked. There was a third problem involving repeatedly subtracting the smaller of two numbers from the larger one. We discussed several possible approaches, but I couldn't identify the logarithmic solution quickly enough. Since the time was nearly over, we left it at the discussion level. I received confirmation that I had cleared the first technical round and was moved to the second one. This round started with a binary-tree problem involving the top view of a tree. I was comfortable with it and coded the solution fairly quickly. The difficult part came next. I was asked to design and implement an LFU Cache. I knew the basic idea and discussed different possible approaches, but I could not arrive at the optimal implementation within the interview. I eventually wrote a less efficient version, but the interviewer was not satisfied with it. With some time remaining, the interviewer asked me basic OS and DBMS questions. I answered those reasonably well, but the cache problem had already affected the round. A couple of days later, I received the rejection email. The experience made one thing very clear to me: for Amazon's SDE interviews, getting to a working solution is not always enough. The interviewer may continue pushing for the optimal approach, and I needed to be much stronger at recognizing advanced data-structure patterns under time pressure.