Question: Tell me about your current job?
Answer: The candidate works on a data management platform for real-time data collection and online activities. Medium and large companies use the platform for online marketing.

Q: Why are you interested in Amazon?
A: She wants to move here because her husband found a job in this city. She is interested in Amazon because it deals with vast amounts of data. She is interested in ML. She enjoys working on distributed systems. She wants to do something new.

Question: Tell me about a time your hard work served the business of your company.
Answer: The candidate described an issue with the data management platform, which allows users to issue HTTP requests to poll data. Each poll produces read and writes. Some of those requests are malicious; they cause a problem for the DB because the SLAs go down. The candidate created a mechanism to detect malicious access patterns.

Q: Malicious in which sense? Fraud?
A: Malicious in the sense of load problems that ruin the weekends of on-calls.
The candidate proposed a cleaner solution, which took more time to implement. It was based on a probabilistic data structure using blooming (?) filters and hash functions to determine the minimal frequent of ids with malicious access patterns.

Comment: Despite asking the candidate, she did not provide data points for the technical details, which would have helped to understand the situation better.
A: The technical lead was opposed to the proposed solution because of concerns about the time it would take to implement the solution. The candidate took “garage time” to implement a POC, which convinced the tech lead. The solution was implemented and rolled out.

Q: What exactly was the business impact? How did you measure it?
A: We had no longer increased latency. There was an SLA of 150 ms for T90 or T95. With the malicious access patterns, we had actual latencies of 200 ms or even 1 s. Wit the solution, we did not have the spikes anymore caused by the malicious access patterns.

Q: How exactly did you measure this?
A: Via incidents. During the introduction of the solution, we had only one incident with a latency between 150 ms and 1s. It might have been 3-4 incidents per year before that.
Comment: This doesn’t really show the business but more the technical impact.

Analysis

Competency asserted: Have Backbone; Disagree and Commit
Job title: Software Development Engineer (SDE II)
Interviewer role: Software Development Manager

Vote: Mixed

The candidate did not provide convincing data points for the business gain, which resulted from her solution to a problem.

Learnings

Software Engineers don’t always have exposure to customers and business figures. When preparing your stories, have one story with a measurable business impact. If you are still working for a company, ask your manager, salespeople, PMs, for figures. Ask them how they evaluate the business impact.

Leave a Reply