Heap

About 50% of our team works on infrastructure, 25% on data ingestion/related problems, and 25% on product work.

We use Requests For Comments (RFCs) as a way to communicate about big projects or propose process changes. RFCs are a great way for us to work asynchronously because they allow people to comment and discuss in their own time, and with greater depth than would be possible in meetings. RFCs also serve as a way of documenting how a feature was built and why it was built.

Check out our engineering blog here: https://heap.engineering/

Data Capture — We want to provide our users with the richest possible dataset. That means building new integrations (like Android), capturing more semantic data (like A/B test groups), and pulling in data from other sources (like Stripe). All of this has to happen with virtually no performance cost or integration overhead.

Real-time Infrastructure — We support an expressive set of queries that allow our users to slice and dice their data in arbitrary ways. The results need to come back with sub-second latencies and reflect up-to-the-minute data. We're scaling an in-house distributed system to make this possible.

Data Visualization — Building Heap is as much a design problem as anything else. The most powerful analytics infrastructure in the world doesn't matter if our users can't access that power and understand the results. We’re iterating on a general-purpose analysis interface that’s easy for non-technical users, but flexible for advanced users.