Revenue recognition accounting. (Is that not interesting to you? It's super interesting to me!) There's this whole world of rules that accountants follow to pr

Revenue recognition accounting. (Is that not interesting to you? It's super interesting to me!) There's this whole world of rules that accountants follow to properly record revenue. These rules are different for physical products, subscriptions, contracts, and even rules for what to do when you bundle these things together. And yet, the only two options for companies to crunch all this data is to use an Oracle product or Excel. (Or get the eng team to write SQL queries, to try and produce a correct result for a topic they don't have a background in!) To do this, we're really leaning into databases like Clickhouse and DuckDB - taking raw data and very quickly summarizing them into journal entries that you can post during the month. Also using kinesis to let users stream in financial data in realtime, so we can then process that data in batch. (Kinesis -> S3 -> Lambda -> DB). In doing this, have also been writing low-level C and Go code to quickly iterate over a stream of transactions and accumulate results.