Pet Dating App

The challenge

How I approached it

• →Swipe-to-Match Engine: Built a gesture-driven card stack using React Native Reanimated with pan gestures supporting three directions—right to like, left to pass, up to super like—with spring physics (damping 26–28, stiffness 320–360), velocity-aware fly-off animations, ±7° rotation, and stacked background cards with progressive scaling.
• →Real-time Chat via Socket.IO: Engineered a full-duplex messaging system using Socket.IO with Redis adapter for horizontal scaling. Supports text, image, GIF, and audio messages with idempotency keys, typing indicators (5s timeout), online/offline presence tracking via Redis, read receipts, message reactions, edits, deletions, and rate limiting (30 msgs/min per user).
• →Voice Message System: Implemented WhatsApp-style voice recording using expo-audio with live animated waveform bars (28 bars), slide-left-to-cancel, slide-up-to-lock hands-free, 800ms minimum / 60s maximum duration, and playback with seekable 40-bar waveform, speed control (1x/1.5x/2x), and haptic feedback.
• →Social Media Feed: Built a complete social layer with post creation (text, images, videos, GIFs, up to 10 hashtags), nested threaded comments with visual connectors, likes, bookmarks, sharing, and real-time new-post notifications via Socket.IO that surface a 'New posts' pill without jarring scroll resets. Powered by Shopify FlashList for buttery-smooth virtual scrolling.
• →Google Maps Integration: Integrated react-native-maps with Google provider and expo-location for GPS-based discovery. Built a backend proxy to Google Places API (key never exposed client-side) showing nearby parks, pet stores, vet clinics, grooming salons, and pet-friendly restaurants with ratings, photos, and deep links to Google Maps for directions. Results cached 15 minutes in AsyncStorage.
• →Location-Based Discovery: Implemented distance-filtered pet suggestions using GPS coordinates, city picker with search and current-location detection, reverse geocoding for city/region names, and privacy controls (hide location toggle, location obfuscation on profiles).
• →Blind Match Mode: Built an anonymous matchmaking feature with Redis-backed queuing (120s TTL), shared icebreaker questions, mutual accept/pass decisions, and automatic match creation on mutual acceptance—all with content filtering and block-check enforcement.
• →Interceptive Middleware & Tier System: Developed a FeatureGate service that intercepts every API request to validate permissions and quotas (swipes/day, super likes, boosts, rewind, read receipts, advanced filters) via Redis-cached tier configurations across Free, Premium, and Tovy Pro tiers.