PRODUCT DESIGNER. FOR COMPLEX SYSTEMS

Challenge: I designed a unified patient-flow system across five surfaces — self-service kiosk, TV queue displays, mobile, speaker and an admin console. The real brief was emotional, not technical: reduce patient anxiety in the waiting room, not just manage token numbers.

Approach: I mapped patient anxiety and waiting-room behaviour through contextual observation, designed token booking, kiosk UI, queue displays and mobile as one cross-device journey, built the admin control dashboard for real-time queue overrides, progressively revealed information to lower cognitive load at each step, and kept voice announcements in strict parity with the on-screen state.

Outcome: Shipped in 5 weeks; live in production. Patients always know where they stand — staff control the flow instead of fighting it.

Link: Read full case study ↗

Challenge: Event guests lose their photos. SnapShare needed an AI platform where a guest at a 2,000-person conference finds every photo of themselves in under 30 seconds — no app download, no account.

Approach: I designed zero-friction guest face-search flows where a selfie is the only input, built photographer dashboards for bulk upload and event management, designed organiser tools for branded galleries and white-label theming, created real-time notification/delivery states, and designed every AI loading, processing, and error state.

Outcome: Live commercial product. Deployed at Huddle Global 2025 and The Summit of Future 2025.

Link: Visit snapshare.ai ↗

Challenge: I designed a compliance platform for one of the world's largest energy companies, serving five user types — contractors, trainers, approvers, HR and management. Scope spanned HSE training approvals, incident reporting, access passes, contractor ID workflows, resort bookings and food ordering — all in one product.

Approach: I mapped all five user roles through stakeholder interviews, restructured the information architecture across 8+ modules, cut average task steps ~40% with role-specific paths, made approval states and status visible end-to-end, and used progressive disclosure to tame dense compliance forms.

Outcome: Each role now sees only what it needs. Stakeholders reported significantly faster onboarding for new users. Visuals are available on request.

Challenge: I designed a ChatGPT-class assistant plus departmental file management inside Aramco's strict security perimeter — every design decision carried data-access implications. Users needed real AI capability without ever seeing a document outside their clearance.

Approach: I mapped the permission model before any screens, designed conversational flows fit for enterprise constraints, designed every AI state (idle, streaming, error, access denied), built permission-aware document preview patterns, and communicated data boundaries to users in plain language.

Outcome: An AI experience that respects enterprise security without feeling locked down — including a deliberately designed denied-access state. Visuals are available on request.

Challenge: Clinical staff were under-reporting incidents because the existing system was too slow to complete under pressure — and a missed incident report has real patient-safety consequences. Scope covered policy management, incident reporting and KPI indicator tracking.

Approach: I ran contextual research with nursing, admin and quality teams, designed progressive disclosure forms to surface only the fields that matter, implemented smart defaults from department and shift context, built KPI dashboards for quality-leadership review, and made document and policy versioning unambiguous.

Outcome: A nurse can file a complete incident report from memory in under 5 minutes. Reporting friction was successfully removed for the people with the least time.

Link: View on Behance ↗

Challenge: I designed one coherent product for three completely different users — customers booking care, providers executing jobs, and admins overseeing everything — without building three separate products.

Approach: I service-blueprinted the full three-sided ecosystem, designed role-isolated navigation and contextual task paths, created booking, payment routing and job-assignment flows, built one notification system shared across all three roles, and designed settlement and admin reporting interfaces.

Outcome: Three distinct interfaces, one design language — each side of the marketplace gets workflows shaped to its job.