In the chance I get an interview I wanted to make a mock protoype of a backend that adds transactions from a geth mempool to a redis stream. I know your architecture is much more complex but wanted to get into the headspace of some of the problems your architecture deals with. Also I realize your products have a heavy API/SDK component as well which I have some experience with on my PayLater Project I am working on for EthOnline (add link).
This code does not in really anyway reflect the quality of code I would put into production. I timeboxed myself to only a 4 hours to do this since I was also competing in Eth Global's hackathon EthOnline, but I had never worked with Redis (have Kafka experience) and have never queried the mempool/txpool directly before so that took up the bulk of the prototype time. I would love the opportunity to dive deeper.
Also I am interested in how Blocknative will approach the coming Roll Up dominant + eth2.0 landscape. It seems Eth2.0 wont fundementally disrupt mempool analysis but ZK roll ups could make analysis much harder. I think the switch to Optimistic execution will only increase the desire for your services.
Thanks for reading!