Danesh – Swift, SwiftUI, Unit testing

iOS education app suite combining offline past-paper practice with an on-device AI tutor. Evolved from a Cambridge exam past-paper browser into a full study platform with a mark-scheme-aware AI tutor that delivers personalized feedback without cloud connectivity. Built with SwiftUI and MVVM, with a Python-based data pipeline processing CIE PDFs and mark schemes into structured, tutor-ready content.

• Repositioned the product from a past-paper utility into a full study platform, connecting content browsing with an AI-powered tutoring layer.
• Led native iOS architecture around SwiftUI, MVVM, explicit stores, and repository boundaries to support testability and long-term maintainability.
• Designed a mark-scheme-first AI tutor system that delivered personalized feedback under real mobile constraints without relying on cloud connectivity.
• Built a supporting data pipeline that processed inconsistent CIE PDFs, mark schemes, and syllabus content into structured, tutor-ready data.
• Improved launch readiness by separating startup work from debugging overhead and reducing resource-package noise.
• Documented an extensible education-app blueprint — covering onboarding, study planning, and content delivery — for future reuse.

Native iOS migration for a large-scale sports betting platform, transitioning from React Native and legacy UIKit to native Swift and SwiftUI across four consumer-facing products. Focused on rendering performance, domain safety around GraphQL-backed models, and architectural consistency across a high-traffic product family. Achieved approximately 233% improvement in initial load time through targeted SwiftUI view invalidation reduction.

• Led native Swift and SwiftUI delivery across high-visibility home, quick-link, BigBets and hybrid widget surfaces, turning React Native-era behavior into reusable native patterns that could ship safely across brands.
• Contributed to a roughly 233% improvement in initial load time by reducing SwiftUI view invalidation, layout negotiation issues and expensive render paths rather than treating performance as a late polish pass.
• Strengthened domain safety around GraphQL-backed product models by keeping identifiers, bet payloads and mapping boundaries typed and reviewable instead of letting backend contracts leak as loosely shaped data.
• Improved reliability in navigation-heavy flows by tightening deep-link handling, feature-flagged routing and state ownership across reusable view models and brand-specific surfaces.
• Raised engineering quality through architecture reviews, ViewModel tests, snapshot coverage and mentoring for engineers moving from React Native into SwiftUI, especially around view identity, async work and state boundaries.

Safety-critical iOS product for field workers where reliability under stress mattered more than feature velocity. Covered emergency alerting, real-time device monitoring, AI-powered SafeSense detection via Vision framework, and check-in workflows. Built with SwiftUI and MQTT-based real-time communication, with native architecture designed to handle push notifications, deep links, and emergency workflows with zero tolerance for failure.

• Set the native iOS architecture for safety-critical workflows, including coordinator-led navigation, typed routes, auth gates and state boundaries that could survive monitoring, alert and permission transitions.
• Designed a single deep-link and notification entry model so push taps, universal links, emergency workflows and in-app routes resolved predictably without duplicating navigation or replaying unsafe actions.
• Built reusable SwiftUI product patterns for emergency actions, status surfaces, check-ins, device state, geofences, permission prompts and media controls, reducing one-off UI while keeping high-stakes states readable.
• Stabilized realtime monitoring by tightening MQTT event handling, cancellation, stale-event behavior and SwiftUI identity ownership so location, device and media changes updated the right surfaces without rebuilding unrelated screens.
• Integrated location, AVFoundation and SafeSense AI/Vision surfaces in a way that kept product behavior responsive while respecting native iOS lifecycle, privacy and reliability constraints.
• Reviewed architecture and Swift code for the risks that matter in safety software: retain cycles, broad dependency lifetimes, uncancelled async work, ambiguous ownership and UI state that could fall out of sync under pressure.

iOS field-auditing platform for energy auditors working across remote sites with limited connectivity. Handled offline data capture, complex audit workflows, and field reporting backed by Core Data and Firebase sync. Automated audit workflows improved auditor efficiency by approximately 8x while reducing data-collection time by roughly 60%.

• Architected the iOS app with SwiftUI, Combine and MVVM-C, giving complex auditing flows explicit state ownership, testable ViewModels and predictable navigation.
• Delivered automated audit workflows that improved auditor efficiency by roughly 8x and reduced data-loss risk by about 70%.
• Built persistence and sync behavior around Core Data, Firebase and Crashlytics so field data could survive intermittent connectivity, user interruption and failed uploads.
• Turned repeated audit steps into reusable components and workflow primitives, reducing the cost of adding new inspection paths over time.
• Established GitHub Actions and TestFlight delivery flows, cutting deployment time by roughly 50% while keeping releases traceable for field operations.
• Led code reviews and architecture discussions that reduced bug density by roughly 25%, with emphasis on state management, data integrity and failure-mode handling.
• Managed delivery with product, design and operations stakeholders through Agile planning, Monday.com tracking and Notion documentation, keeping technical tradeoffs visible rather than hidden in implementation details.

iOS energy-auditing app for field auditors across the United States, built on a hybrid UIKit/SwiftUI architecture to modernize core workflows while preserving reliability in the existing programmatic codebase. Supported structured data capture, CI/CD-driven releases, and audit workflow extensions without requiring a full rewrite.

• Built and modernized core audit workflows using a hybrid UIKit/SwiftUI architecture, preserving reliability in the existing programmatic codebase.
• Replaced manual paperwork with structured forms, guided checklists, and offline data capture tools.
• Set up AppCenter and GitHub-based CI/CD, reducing production release time and improving release reliability.
• Reviewed pull requests with a focus on correctness, maintainability, and regression risk.
• Managed delivery with product and field stakeholders through Agile planning and Trello/Shortcut tracking.
• Maintained an extensible architecture so new reporting requirements and audit variants could be added without full rewrites.

iOS onboarding work on a major fintech platform's New User Experience (NUX) flow, focused on test coverage, accessibility compliance, and safe delivery in a large-scale mature codebase. The entry path affected millions of users, requiring intentionally narrow changes balanced with measurable quality improvements.

• Implemented the first meaningful test coverage around the NUX onboarding area, creating a safer baseline for ongoing feature delivery.
• Improved accessibility coverage by approximately 15%, aligning onboarding behavior with platform usability standards.
• Delivered UIKit features within CLEAN architecture, keeping presentation, domain logic, and integration layers separate.
• Used Jenkins, Fastlane, Cucumber, and Appium to improve automated test execution and reduce CI flakiness.
• Used Jira and Confluence to keep delivery state, blockers, and technical decisions visible across distributed teams.
• Kept changes intentionally narrow in a mature codebase, avoiding unnecessary rewrites that could introduce regression risk.

iOS social music-discovery app focused on browsing, sharing, and discovering songs. Performance-sensitive feeds were built with AsyncDisplayKit/Texture to handle real-world scrolling conditions, with AVKit-backed playback state managed within an MVVM architecture. Shipped in a startup environment with investor-facing milestones.

• Led iOS delivery and team execution for the core music-discovery experience, balancing startup speed with code quality.
• Built performance-sensitive feeds using AsyncDisplayKit, reducing scrolling latency under real device conditions.
• Integrated AVKit playback and media handling within MVVM, keeping playback state separate from UI rendering.
• Supported GCP-hosted backend integration and worked through API contracts for content delivery features.
• Improved release workflows with GitHub Actions and Fastlane, reducing release-cycle overhead by approximately 40%.
• Managed cross-functional sprint delivery through Trello and Monday.com.
• Contributed to investor-facing demos by explaining performance work, product architecture, and roadmap decisions.