Building the data layer behind real-time analytics
Built real-time analytics and e-commerce attribution APIs around modular services, explicit validation, and high-volume query performance.
- Role
- Full-stack development, backend architecture
- Scope
- TypeScript, Elysia, ClickHouse, query builders, data executors
- Year
- 2025–Present
Analytics systems are only useful when their answers can be trusted. The work here spans the layers between a product question and the data that answers it: service boundaries, request validation, query construction, and execution against high-volume event data.
I worked toward a modular architecture in which those responsibilities stay explicit. That makes it easier to add a report or attribution dimension without quietly changing the meaning of an existing one.
The satisfying outcome is not a more complicated data layer. It is an analytics product that can answer more useful questions while keeping scope, performance, and correctness visible.