Selected Work

Problem

Pet owners often juggle medication schedules, vet visits, walks and photos across scattered notes and apps, leading to missed doses, forgotten appointments and lost memories.

What I did

• Designed and built a native SwiftUI app from scratch for iPhone, Mac and Apple Vision from a single codebase
• Built multi-pet profile management with single-tap switching between animals
• Implemented smart reminders for medications, vaccinations, deworming and vet visits, plus a timeline of health records and life events
• Added a pet safety module with lost-pet alerts and a map of nearby pet care facilities
• Built a premium tier with shareable pet profiles (view-only or full-edit access) and recurring events
• Owned the product end-to-end - design, development, App Store release and ongoing updates

Result

• Live on the App Store with a 4.7/5 rating(approximate, shared within NDA limits)
• Available on iPhone, Mac and Apple Vision from one SwiftUI codebase
• Free app with a premium tier for family sharing and recurring events

Problem

Internal banking tools relied on long, multi-step forms with strict regulatory and validation requirements, built inside a security-restricted environment (no external scripts, strict access and audit controls). Edge cases were mostly caught manually, and validation bugs occasionally reached production.

What I did

• Owned frontend implementation of validation-heavy, multi-step forms and critical user flows in React/Next.js
• Built reusable, schema-based validation and form-state patterns shared across multiple flows and teams
• Worked within a security-restricted banking environment - strict access controls, audit and compliance requirements
• Partnered closely with backend, QA and security to keep flows compliant end-to-end

Result

• ~30% fewer validation-related bugs reaching production after introducing shared, schema-based validation components(approximate, shared within NDA limits)
• ~25% faster implementation of new multi-step forms thanks to reusable field and validation building blocks(approximate, shared within NDA limits)
• 10+ multi-step flows delivered and maintained, used daily by thousands of users(approximate, shared within NDA limits)

Problem

As the number of critical user flows grew, regressions increasingly slipped into releases. Manual regression testing before each release was slow, inconsistent, and didn't scale with the size of the app.

What I did

• Introduced and extended a Playwright E2E test suite covering the app's critical, highest-risk user flows
• Wrote resilient, maintainable tests (page objects, fixtures) to keep the suite stable as the UI evolved
• Integrated the suite into the team's CI pipeline, running automatically on every pull request
• Worked with QA to prioritize coverage of flows with the highest regression risk

Result

• 15-20 critical flows covered by automated E2E tests(approximate, shared within NDA limits)
• ~40% fewer regressions reaching production after rollout(approximate, shared within NDA limits)
• Several hours of manual regression testing saved per release cycle(approximate, shared within NDA limits)

Problem

The client's storefront - a large B2B/B2C catalog of office and institutional furniture - had slow page loads and a non-native mobile experience, which hurt engagement across a large, image-heavy product range.

What I did

• Built the customer-facing frontend in React/Next.js integrated with a Magento backend
• Implemented PWA features - offline support, installability, faster repeat-visit loads
• Focused on Core Web Vitals and rendering performance across a large, image-heavy product catalog
• Worked directly with design and backend teams to ship incrementally without disrupting an active storefront

Result

• Lighthouse performance score improved from roughly ~55 to ~85+(approximate, shared within NDA limits)
• ~40% faster load time on key catalog and product pages(approximate, shared within NDA limits)
• Noticeable increase in mobile engagement after the PWA rollout(approximate, shared within NDA limits)