I interviewed with the Agentic AI security product team under AWS. The location is Austin.
Every round was 1 hour.
Round 1 was a DAG topological sort. It was pretty good, pretty basic.
Round 2 was another directed graph problem. This team really likes graph problems. It was something like working out what you need to make a cake, and it is very similar to a LeetCode problem.
Round 3 was a traditional system design: design an AI agent, tech, security stream, processing security detection services, with 1TB per day of traffic.
The second half hour of every round was 2 BQs, all the common Amazon LPs.
Discussion
Loading comments…