About algorithms: It very much depends on the thing one works on, whether one ever needs to look up what a good algorithm is for a problem, which needs solving. Often it is more about how to put pieces together, using APIs, building an API, making a service, etc. Similar for data structures. For most you only need to know the basics like when to use a hash table, when to use an array or whatever of the data structures that are often available anyway in the standard library of the language.

This even more diminishes the gains that engineers get from their expert knowledge.

Usually only when in a position to develop something ground up, does the expert knowledge of data structures and algorithms come into play.

Algorithm and data structure questionaires rather seem like mathematics in many cases. A filter to weed out people, but not often of use at the job.

Obviously I might be biased due to personal experience.