Product portfolio

Product thinker for AI, mobility, commerce, and operational workflows.

This portfolio is organized for PM evaluation: user problem, metric, personal ownership, product tradeoff, system depth, evidence, and interview-ready follow-up for each project.

3 domainsTransit, tourism, and retail checkout show range across operational and consumer workflows.
0.35 secSMART COMMUTE beat its 10-second route recalculation target in verification testing.
51M+WESAL cites integration into Nusuk; add source link before sending to recruiters.
25k unitsPaperless cites a Foodics co-sell channel; add proof artifact or reference.

Selected work

Case studies structured around PM signal.

Each project now separates what the candidate did, why the decision mattered, what evidence exists, and what an interviewer should probe.

01Public transit routing

SMART COMMUTE

Demand-responsive bus routing for Dhahran and Al Khobar

Transit opsVRPTWComputer visionHeuristics

Turned static bus scheduling into a live-demand routing prototype for transit operators managing wait time, crowding, and fleet efficiency.

Role
Optimization lead for routing and scheduling
Users
Commuters, bus operators, dispatch/admin teams, and riders with accessibility-sensitive service needs.
Product question
How might a transit operator adapt routes fast enough to improve rider experience without creating unreliable operations?

Personal ownership

  • Set the metric hierarchy around rider wait time, crowding, information accuracy, operating cost, emissions, and service guarantees.
  • Translated product goals into routing requirements: fast recalculation, delay limits, route stability, and accessibility constraints.
  • Designed the two-phase heuristic, Cheapest Insertion plus Simulated Annealing, around an under-10-second operating requirement.

Key decisions

  • Prioritized service reliability over pure solver optimality because riders and operators need updates they can trust in real time.
  • Benchmarked the heuristic against exact solvers and Solomon VRPTW instances to separate speed gains from untested model behavior.
  • Connected stop cameras, bus GPS, maps, security, backend services, image-processing inputs, optimization, and database storage into one operating path.

Evidence

  • Architecture diagram available.
  • Route map artifact available.
  • Verification results available: 0.35-second average route recalculation time, 12.4-minute max wait time, 100% accessible seat availability, 18.5% fleet capacity utilization, and 8.0% delayed passengers over 15 minutes.

Interview probe

Tradeoff to discuss: when should the product preserve route stability instead of rerouting to chase a local optimum?

SMART COMMUTE system architecture diagram showing commuter and admin apps, bus stop computer vision cameras, bus GPS data, map services, private cloud security, backend services, image-processing parameter generation, optimization models, two-phase heuristics, and database storage.
System diagram: live demand, GPS, and map data move through private cloud security into backend services, image-processing parameter generation, and the optimization engine.
SMART COMMUTE route map showing colored bus routes and station markers across Dhahran and Al Khobar.
Route map: candidate routes and station clusters across Dhahran and Al Khobar.

Verification results

ISE verification matrix

PASS0.35 seconds avg
Route recalculation time

Target: 10 seconds

PASS12.4 minutes
Max wait time

Target: 30 minutes

PASS100%
Accessible seat availability

Target: 90%+

PASS18.5%
Fleet capacity utilization

Target: 10%+

PASS8.0%
Delayed passengers over 15 min

Target: 33% benchmark

02Tourism trip planning

WESAL

Trip planning across transportation, lodging, destination access, and routing

Agentic AITourismMarketplacePersonalization

Contributed to a patent-pending tourism platform that turns one travel request into a coordinated itinerary across multiple providers.

Role
Product workflow and agent-architecture contributor
Users
Travelers, travel agencies, transport providers, hotels, destination-access services, and marketplace operators.
Product question
How might a traveler express one intent while the platform handles provider search, access requirements, pricing, and itinerary assembly?

Personal ownership

  • Collapsed separate planning steps into one text or voice request.
  • Designed itinerary logic around budget, party size, walking distance, provider availability, and destination access requirements.
  • Mapped specialized agents for trip assembly, service bundling, marketplace search, pricing, billing, payment, and supply-demand matching.

Key decisions

  • Used triage and orchestration rather than a single broad agent so different workflow stages could be owned, evaluated, and improved separately.
  • Kept traveler profiles, provider data, analytics, and saved preferences in the loop so recommendations stay tied to context.
  • Connected provider paths for flights, intercity rail, hotels, and Nusuk services instead of treating itinerary generation as a standalone chatbot.

Evidence

  • Architecture diagram available.
  • 1st place in an international travel services challenge is claimed and should be linked.
  • Integration into Nusuk, described as a 51M+ user platform across 190+ countries, is claimed and should be linked.

Interview probe

Tradeoff to discuss: how should the product balance traveler personalization, provider incentives, price transparency, and fulfillment reliability?

WESAL multi-agent architecture diagram showing users and travel agencies interacting with a WESAL agent connected to triage and orchestration, trip, bundling, marketplace, pricing, billing and payment, supply-demand matching, data collection, analytics, user profiling, personalized cloud profiles, and external data sources.
System diagram: a triage and orchestration layer coordinates specialized agents for trip assembly, service bundling, marketplace options, pricing, payment, supply-demand matching, analytics, and user profiling.
03Retail checkout

Paperless

One-tap payment, loyalty, and invoicing without asking for a phone number

RetailPaymentsLoyaltyPrivacy

Invented a checkout method that removes phone-number handoffs and paper receipts while preserving retailer needs around payment, loyalty, invoicing, and records.

Role
Product inventor
Users
In-store shoppers, cashiers, retail operators, loyalty teams, and payment or POS partners.
Product question
How might checkout capture payment, loyalty, and invoice data without making shoppers disclose personal information at the counter?

Personal ownership

  • Reframed receipt and loyalty capture as a checkout privacy problem, not a receipt-storage problem.
  • Designed the flow to work in one tap without requiring an app download, QR code scan, or phone-number exchange.
  • Aligned shopper convenience with retailer needs for loyalty identification, payment completion, invoicing, and receipt records.

Key decisions

  • Made privacy and checkout speed the wedge rather than building another post-purchase receipt organizer.
  • Kept cashier workflow simple because adoption depends on reducing friction for both the shopper and the store.
  • Positioned distribution through an existing POS channel instead of relying only on direct merchant sales.

Evidence

  • Two patents pending are claimed and should be linked when public.
  • Foodics co-sell partnership for 25k units is claimed and should be supported with a public or confidential interview artifact.
  • No implementation artifact is currently shown; a flow diagram or demo would materially strengthen this project.

Interview probe

Tradeoff to discuss: what adoption metric proves this is a better checkout behavior for shoppers, cashiers, and merchants at the same time?

Product signal matrix

What this portfolio should prove in a hiring loop.

Product judgment

Frames each project around a specific friction: rider wait time, fragmented trip planning, and checkout privacy.

Analytical rigor

Defines metric hierarchies, operating constraints, routing benchmarks, and adoption proof before discussing the stack.

Execution range

Work spans infrastructure-heavy transit operations, multi-provider tourism planning, and retail payment workflows.

Technical fluency

Can reason across optimization, computer vision inputs, agent orchestration, billing, data systems, and mobile/web surfaces.

Communication

Explains technical choices in user and business terms so nontechnical and technical reviewers can both evaluate the work.

Remaining proof gaps

Only unresolved claims remain here.

These items still need real source links or shareable artifacts before the portfolio is fully recruiter-ready.

  1. Add a public source for the WESAL challenge result and Nusuk integration, or decide what confidential wording to use.
  2. Add a WESAL patent reference if it is public or shareable.
  3. Add public or confidential proof for the Paperless patent claims and Foodics co-sell partnership.
  4. Add a Paperless checkout flow artifact or demo screenshot.

Candidate details

Abdulmuhsen (Abdul) Fakih

MS, Business Analytics at Cornell University, Aug 2026 to May 2027. Based in New York City, NY.