Keeper Tax

We have 7 engineers, including myself (CTO). Everyone is a full-stack generalist, and, generally speaking, when you ship a feature, you will own it end-to-end (frontend and backend). We meet daily for standup to talk about what we're working on. We also have weekly and monthly checkins for planning purposes - I want to evaluate our progress at a high level, so we can make adjustments if needed in order to achieve our long term plan.

Here's one example: To implement tax filing, we're building on top of an unreliable, relatively slow API. This means that, for our product to appear fast to users, we need to create background jobs, and we need to have retries for those jobs. The retry framework is working fairly well, but I think there are a lot of things we can do to improve it.

Here's another example: We currently have a background job that processes transactions for all users, but this requires a huge amount of memory. On our roadmap, we'd like to break this job into smaller batches of users, and run those jobs in parallel. This should both reduce memory usage and increase performance. The code is a bit complicated to break up right now, but it's something we'd like to do in the future.