Abdul Fakih

Portfolio

I turn multi-party workflows into single actions: one tap to checkout, one request to a booked trip, one live signal to reroute a bus.

Cornell MS Business Analytics candidate building products across software and hardware operations: demand-based transit routing, multi-provider travel planning, and NFC checkout at the POS.

View selected work Email Abdul

Selected Work

Retail Checkout

Checkout today is a chain of separate steps: pay at the terminal, scan a QR code or type your email for loyalty, ask for an invoice, then wait for a receipt. We built a one-tap NFC checkout that plugs into the existing POS flow and completes all four at once. No app, no QR code, and no personal details to hand over.

  • 2 patents pendingMethod and system protection for the checkout flow.
  • Foodics co-sell: 12 live pilot terminalsSigned channel partnership; 25k-terminal network.

Selected for the NVIDIA Inception Program and Recognized by the NFC Forum

Judgment

A receipt app misses the real problem: checkout friction comes from extra steps, and every app download or QR scan adds one. The better answer is a single NFC tap built into POS and SoftPOS flows, using the wallet shoppers already carry, such as Apple Pay. One tap completes payment, loyalty, invoicing, and receipts. Cashiers keep their workflow, retailers keep their records, and shoppers install nothing and share no personal data.

The trade-off is forgoing a consumer app of our own, in exchange for adoption that requires no change in shopper habits.

Role
Product Owner
Problem
Separate payment, loyalty, invoice, and receipt steps slow checkout and demand extra apps, scans, or personal data.

Travel Planning

Led product workflow and agent architecture for a patent-pending travel platform that turns one request into a coordinated trip across transportation, lodging, destination access, pricing, and payment.

  • 1st placeInternational competition with 150+ finalists from 20+ countries.
  • Nusuk integration (an app used by 51M people in 190 countries)Live integration.

Judgment

The risk was generating a plan that could not be booked or priced, and a planner that suggests anything looks impressive in a demo but fails at checkout. So the workflow separates intent capture, provider availability, pricing, payment, and profile memory, and it checks availability and price before a recommendation reaches the traveler. Every itinerary shown is one they can actually pay for, and providers receive bookings instead of leads.

The trade-off is forgoing open-ended destination inspiration, in exchange for plans that survive checkout.

Role
Optimization & Front-end Prototyping
Problem
Disconnected travel services prevent coordinated plans with confirmed availability, pricing, and payment.

Public Transit Routing

Led routing and scheduling for a demand-responsive bus prototype using CCTV-based demand sensing to reduce wait time, crowding, and wasted fleet capacity.

  • 12.4 min max wait timePrototype result: 32% better than current static scheduling.
  • < 8% delayed ridersPrototype result: riders delayed over 15 minutes, below the 33% ceiling.
  • CCTV + bus GPS inputsStop demand and live vehicle location feeding routing decisions.

Judgment

The priority was less rider waiting, not solver elegance. A route that is optimal on paper but changes every few seconds makes service unpredictable for drivers, dispatchers, and riders with accessibility needs, so routing updates are bounded by dispatch control, accessibility, and reliability constraints. CCTV is used only to count riders at each stop, not to identify them, so demand reaches the router as a number, not as footage.

The trade-off is accepting slightly less optimal routes, in exchange for a service operators can trust and riders can predict.

Role
Team Leader
Problem
Reduce waiting, crowding, and delays without making live dispatch unpredictable.

My Approach

Start with the step that's failing, not the feature.

Design for the transaction, not the recommendation.

Bound the system so live operations stay predictable.

Fit the behavior people already have.

Contact

Abdul Fakih

New York City, NY.

af729@cornell.eduLinkedIn