Case studies

Application Problems, Engineering Decisions, Delivered Changes

Selected work across React interfaces, PostgreSQL workflows and application reliability.

Featured study

Lamiya Wedding Center · 2026 · E-commerce

13 min read • 2470 words

Implemented optimistic cart updates with TanStack Query, snapshot-based rollbacks, and a multi-stage checkout connected to a PostgreSQL reservation backend.

Optimistic cart updates with rollback and server reconciliation.

  • Optimistic cart updates before server confirmation.
  • Snapshot-based rollback when a mutation fails.

React · TanStack Query · Optimistic UI · E-commerce Frontend

Read case study

More case studies

Locator — personal project · 2026 · Web Application Engineering

Building an Offline Location Journal That Recovers from Interrupted Uploads

Stable retry identities, staged cloud saves, and atomic local acknowledgement implemented.

Offline · IndexedDB · Upload recovery · +3 more

8 min read • 1557 words

Aurora Graphics · 2026 · Marketing

Reducing Form Spam in a Next.js Application

Layered form protection without reCAPTCHA.

Next.js · Supabase · PostgreSQL · +5 more

7 min read • 1275 words

Lamiya Wedding Center · 2026 · E-commerce

Preventing Overselling with PostgreSQL Reservations

Reservation-based inventory handling with idempotent webhooks.

PostgreSQL · Inventory Management · E-commerce Architecture · +5 more

10 min read • 1909 words

Portfolio · 2026 · Website

Improving CMS Autosave and Publishing Reliability

Autosave recovery and publishing status safeguards with an edit history.

Next.js · React · TypeScript · +5 more

14 min read • 2706 words

Engineering methodology

  1. Reproduce the failure

    Trace the request or interaction and capture the conditions that trigger it.

  2. Choose the boundary

    Decide where the guarantee belongs, from a client rollback to a database transaction.

  3. Verify the guarantee

    Test the expected flow alongside cases like concurrent requests and failed writes.

  4. Check production behavior

    Use measurements and regression checks to verify how the change holds up under real workloads.

Have a similar problem in your application?

Share what’s happening and I’ll suggest a practical next step.

Discuss your project