I build products.
And the foundations
they grow on.

The decision

• Moved aggregation from the browser into Elasticsearch aggregations.
• Redesigned the APIs so each component could fetch focused data on its own.
• Added request caching and de-duplication.
• Kept Web Workers for the computation that still belonged on the client.
• Built the organization’s design system and removed legacy dependencies.

The outcome

Reduced typical loading times from 40–60 seconds to approximately 0.5–1 second. Dependency cleanup cut the bundle by about 35%, and the design system helped deliver a major rebrand in approximately 20% of the estimated time.

The decision

• Organized both apps in an Nx monorepo with a shared core of components, hooks and application logic.
• Made React composition the architectural basis: small, semantic primitives that combine into different workflows.
• Introduced stable TypeScript contracts so teams could work in parallel.
• Established unit testing as a standard requirement.

The outcome

Gave both mobile apps a shared foundation for business logic, UI and infrastructure. Engineers could build features from tested, reusable components and apply improvements across both products.

The decision

• Designed an incremental path so architectural migration and new features shipped together.
• Redesigned parts of the architecture and delivery process behind SAML connector creation to remove friction for the whole team.
• Reorganized responsibilities and trained an engineer who had mostly built repetitive connectors, widening what the team could take on.

The outcome

Saved approximately 7–8 months of engineering effort through incremental migration. Improvements to SAML connector development contributed to an approximately 350% increase in connector output over five months.