Waste, Waste, and Waste.

Part 1

It's a complete waste of your time. HRs are pretty immature. After two rounds, which went pretty well, they suddenly asked for salary details and expectations. Three interviews were pending, but after just one additional interview, all remaining interviews were canceled. I quickly called HR; she got nervous and said one round didn't go very well. The head HR would contact me for details and then cut the call. I didn't receive any call after that.

All the interviews went pretty well. The interviewers themselves said they were satisfied with my solutions.

I am sure they realized that they cannot give a hike or meet my expectations. They are just mass interviewing people without looking at their current salary and expectations. That would result in a waste of your precious time.

Part 2

You may be surprised to see that guys with 2-4 years of experience are interviewing guys with 15 years of experience. Interviewers are theoretically and technically good, but they lack some system design aspects because they haven't seen or faced real production design issues and don't have real experience seeing/doing much design work. They are immature in that. So, stick to basic theoretical design if you want to make them happy (that's what I did :)).

1. Given a JSON containing a list of employee objects, perform some aggregation, such as calculating the average salary of employees, in an object-oriented manner.

2. How do you design user impersonation (test user)? The test user should have the same experience as a real user. Also, how do you ensure that the test user has read-only access, not write access?

3. System design: You are receiving events from mobile and other devices (like user click events). Design a system that will retrieve and process these events and support queries such as getting the history of the number of clicks per hour over the last 24 hours. Design for fault tolerance, scalability, HA, low latency, etc. (very theoretical).

Follow-up: How do you design a client SDK that handles the server not being reachable or spikes in the system?