1st round - Leetcode type coding test on putty. The server was in US, and I'm typing from Sydney. So lots of lag. Anyway cleared it. In this day and age where there are very good online code collaboration tools, not sure why they are going with putty. 2nd round - Asked about a multithreading scenario. Answered around 80% of it correctly. Same problem with putty again. 3rd round - Started off innocently with STLs. Then the interviewer started deep diving into unordered_set implementation and asked me to code it up. At one point of time to answer the interviewer's subquestion - I told to make an assumption that a pointer is 4 bytes. He started laughing. I didn't have much hope after that and predictably the interview went south from there.

Pros: Great initial approach – the recruiter thoroughly checked my GitHub profile and projects before the interview, which is very uncommon and shows they do their homework. Cons: A major mismatch between the job description ("Software Engineer C++") and the actual interview reality. The technical stage on CoderPad strictly tests bare-metal C98 skills: raw pointers, manual bitwise operations, and packing bytes into 64-bit integers. If you are accustomed to modern C++ (RAII, templates, safe memory management), this will feel like a massive step backward. Additionally, the time expectations for live low-level bit-shifting were unrealistic, to the point where the interviewer started solving the task themselves. Interview was kind and not stressful, just minor mismatch of naming