Skip to content
All work
Adtech · SaaS platform

Faster decisions on large data sets

Modernising frontend architecture and data flows in a distributed adtech platform spanning multiple business domains.

OpenX

35%+

lower Time to Interactive in reporting dashboards

Project context

The problem that required change

Reporting dashboards processed large advertising data sets, and their usefulness depended on how quickly people could reach a clear view of the metrics.

Several React applications were developed in parallel. They needed shared contracts, components and state patterns without duplicating solutions across domains.

Scope of responsibility

From diagnosis to delivery

  • 01Designing shared architecture for multiple React applications in an Nx monorepo
  • 02Leading the migration of key flows from REST to GraphQL
  • 03Refactoring data fetching and advertising-metric visualisation
  • 04Working with Product, UX and backend teams on API contracts
  • 05Introducing profiling, Lighthouse CI and integration tests into release validation
Process

Decisions organised into a controlled process

  1. 01

    Diagnose data flows

    Profile dashboards and identify where fetching and transformation delayed interaction.

  2. 02

    Shared architecture

    Consolidate GraphQL clients, state and UI components within the Nx monorepo.

  3. 03

    Refactor reporting

    Reduce overfetching and simplify query logic for the most important views.

  4. 04

    Guard against regressions

    Add profiling, Lighthouse CI, Cypress and GitHub Actions to regular release validation.

Measurable outcomes

35%+

lower Time to Interactive on large data sets

1

shared architecture across multiple React apps and business domains

CI/CD

earlier detection of performance and interface regressions

TypeScript · React 18 · Nx · GraphQL · Redux Toolkit · React Query · Cypress · GCP

Next step

Facing a similar challenge? Let’s discuss a concrete outcome.

Share what is currently limiting your product or team. I’ll reply directly with a practical first step.

Discuss your project

Next case study

Automation resilient to advanced protection

Some Say · Aleksander Szmigiel

Facing a difficult technical problem?

Email me