Maqtorah is a two-sided marketplace for moving livestock, equipment, furniture basically anything bulky across Saudi Arabia. Customer posts a shipment, transporters bid on it, delivery gets tracked live. We worked with Ali Aldughairi on both apps: the customer side and the driver side.

A major part of the project involved building RTL support into the user interface. Not just flipping the layout, the actual hard part was mixed-direction content. Order IDs, IBANs, phone numbers, SAR amounts these are Latin strings sitting inside Arabic sentences, and RTL doesn't have a clean answer for that out of the box. We had to work out a rule that held up on a chat screen, a checkout screen, a tracking screen, all differently, and make sure none of them broke when the numbers got longer or the names got longer.

Then there was the density challenge. A single freight request has cargo type, route, receiver info, timing, vehicle needs a lot to carry on a 375px screen. A transporter scanning bids needs to decide in a few seconds whether something's worth their time, and a customer comparing five offers needs to trust the price without having to think too hard about it. Getting that hierarchy right took a few passes.

The two apps also needed to feel like they were built for completely different people. Customer app leans on trust profiles, vehicle photos, ETA, that kind of thing, all visible before you commit. Driver app is the opposite: stripped down for someone driving, checking their phone for two seconds at a time. We ended up with something we called Road Mode, where geofencing picks up most of the status changes automatically and the driver's really just tapping Loaded and Delivered.

By the end we had a full Figma design system for both apps, onboarding through payout, built so it can stretch to new cargo types or new regions later without starting over.

One thing I'm still chewing on: how much should a driver app automate away versus leave in the driver's hands? Curious where others land on that.

Less distraction always works best. Great project Yurii🔥

Wellcue – Comprehensive Health Tracking Mobile App

Wellcue is a production, cross-platform mobile app (iOS and Android) that helps people build and sustain healthy habits by tracking all of their health and wellness data in one place. It's built with Flutter on the frontend and powered by Supabase for the backend — PostgreSQL, authentication, storage, and Edge Functions — and is fully localized in English and Spanish.

Users log and monitor a wide range of health categories: blood pressure readings, exercise activities and duration, nutrition and hydration intake, medication schedules, sleep patterns, and body measurements. Each category supports detailed, timestamped entries and clear progress views, so people can see trends over time and stay consistent.

At the heart of the app is a robust reminder and notification engine. Users create fully customizable reminders with flexible recurring schedules (daily, specific days, or custom intervals) and receive reliable push notifications to keep their routines on track. The system is engineered to handle the real-world edge cases most apps miss: exact-time delivery, timezone-safe scheduling of the next occurrence, deduplication across devices, per-occurrence tracking, and automatic "missed reminder" follow-ups when a scheduled action goes unaddressed. Time-sensitive medication reminders are treated as critical health notifications on iOS.

Engineering highlights:

Reliable scheduled push notifications via a Supabase Edge Function + cron backend (Pushy / FCM), with retry, deduplication, and delivery verification

Strict per-user session isolation, guaranteeing that no data leaks across accounts on shared devices

Subscription and free-trial system integrated with RevenueCat and synced through a server-side webhook

Privacy-first monetization: GDPR/UMP consent flow and non-personalized ads (Google AdMob)

Riverpod state management on a Clean Architecture foundation, with resilient background sync

Full internationalization (EN/ES), including Unicode-safe validation for Latin American input

Wellcue is live and actively maintained on the App Store and Google Play, and meets Apple's and Google's health, privacy, and advertising review requirements.

Flutter with Supabase Edge Functions is a solid pairing here. The part I'd stress-test is a reading logged with no signal, since health data is what users assume is safe the second they hit save. Vitals and medication schedules also tend to want different conflict rules, because...