Dawar Hasnain.

A vision screening app offering four clinical-style tests — visual acuity, color blindness, color contrast, and astigmatism — designed as a triage tool that guides users toward a professional eye examination.

• Solved a core bias in digital vision testing, where users unconsciously move closer to compensate for poor vision, by using ARKit head-position tracking to enforce a consistent device distance throughout each test.
• Mirrored real clinical examination constraints in a mobile context, improving test reliability over unconstrained digital assessments.
• Built structured PDF report generation with PDFKit so users can export results and bring them to a professional eye exam.
• Integrated StoreKit 2 in-app purchases to unlock advanced reports and premium features.
• Followed MVVM and SwiftUI best practices for a scalable, maintainable codebase.

An Augmentative and Alternative Communication (AAC) app helping adults with aphasia regain functional communication through visual, symbol-based interaction.

• Researched the AAC landscape for stroke and aphasia patients, using findings to drive feature decisions around usability and cognitive load.
• Integrated ARASAAC pictograms into a structured, customizable symbol library supporting varying levels of language impairment.
• Implemented on-device ML for predictive word and phrase suggestions, reducing the effort needed to compose messages.
• Built multilingual text-to-speech across four languages with AVSpeechSynthesizer for non-English speakers and caregivers.
• Designed an accessibility-first interface with large touch targets and simplified navigation for motor and cognitive limitations.
• Navigated and resolved an App Store Guideline 2.1 rejection, and shipped localized multilingual App Store descriptions for an international release.

A shift and earnings tracker for delivery drivers — logging per-order earnings, tips, and hours against daily targets, with weekly and monthly performance views.

• Replaced arbitrary income goals with a percentile-based target engine that derives realistic daily targets from the user's own historical earnings distribution.
• Solved overnight shifts crossing midnight by keeping a shift 'active' until its true end time instead of resetting at the calendar day boundary.
• Captured true per-order and per-tip timestamps for hour-of-day and best-start-time analysis, surfaced through Swift Charts (earnings-per-hour by weekday, 8-week trend).
• Built a customizable, direct-manipulation dashboard (drag-to-reorder, resize, rename, show/hide) with independently persisted weekly and monthly layouts.
• Integrated CloudKit private-database sync with a custom mapping layer supporting schema evolution across multiple updates.
• Added a Siri App Intent for hands-free shift logging and a privacy mode that masks earnings while preserving rate-based metrics.

Exam-prep for the Italian patente theory test, built on the official ministry question bank — ~716 concept clusters (Blocchi) and 7,142 true/false questions.

• Designed a daily study system combining concept-by-concept practice with full mock exams, structured around the ministry's own topic taxonomy.
• Engineered a readiness forecaster estimating exam-day preparedness from practice history and weak-question performance for targeted guidance.
• Built a 'Today's Session' daily flow assembling a personalized question queue from progress and weak areas.
• Implemented Supabase auth with Apple Sign-In and a guest 'Skip for Now' flow with milestone-based prompts encouraging account creation.
• Closed an English-speaking-expat usability gap with bilingual question glosses, tap-to-translate Italian lookup, and a dedicated glossary.
• Rebuilt the core learning model from a Duolingo-style vocabulary engine into the current exam-bank-driven model, fully removing the legacy architecture.