Analytics software has a design failure mode everyone recognizes: impressive dashboards full of charts that don't tell anyone anything actionable. Lots of color. Lots of data. No signal. Affiniti was building a SaaS analytics platform that needed to do the harder thing - surface the information that actually matters to the people who need to act on it.
That requires product judgment as much as engineering execution. Building good analytics software means understanding what questions users are actually trying to answer. The technical requirements were real too: a responsive, real-time-capable frontend, a Python data processing layer with real analytical depth, and an architecture that could evolve as the product learned what users needed.