During the interview, though, it felt like the discussion revolved around a very specific, personal interpretation of the pattern. The interviewer seemed fixed on a single flavor of the approach and didn’t really acknowledge or engage with the broader, well-established variants that exist across the industry. Some of the critiques suggested that the design wasn’t fully understood, rather than that the design itself was inherently weak.

Similar issues came up in the architectural follow-ups: several assumptions made during the conversation didn’t match the scenarios I actually described. It felt less like an open technical evaluation and more like checking alignment with one preconceived design style. He hinted that "NoSQL scales better than relational databases" as if that was some universal, unquestionable truth. There was no nuance, no context, and nothing else. Has he even heard of NewSQL or distributed SQL systems that give you horizontal scalability and proper transactions? Or the fact that a lot of scaling problems can be solved with sharding, partitioning, and good schema design on relational databases, instead of just throwing everything into a document store and calling it a day? I know Monzo uses Cassandra, but that's not the best way and that doesn't come without some compromise.

There were questions about previous experience and architectural decisions that were made.