YallaBaytak - Multi-Role Real Estate Operating Platform

YallaBaytak: Multi-Role Real Estate Operating Platform

ad

Problem & System Context

Regional real estate marketplaces often suffer from communication bottlenecks between property seekers, listing agents, and brokerage supervisors. Traditional listing websites act as static classifieds, lacking verification workflows, structured tour scheduling, and immediate notification channels.

I was the sole developer responsible for designing, building, and deploying the complete YallaBaytak platform. It standardizes the property listing lifecycle, automates lead assignment, coordinates walkthrough appointments, and connects agents with buyers through automated WhatsApp event routing.

Architectural & Technical Implementation

1. Multi-Role RBAC & Access Boundaries

The platform manages distinct user roles with strict middleware isolation:

• Client: Browse verified listings, bookmark favorites, submit viewing requests, and track appointment statuses.
• Agent: Create property listings, manage client leads, schedule walkthroughs, and receive instant tour booking alerts.
• Supervisor: Review agent listing submissions, verify property ownership documents, and approve listings for public display.
• Admin: Manage platform analytics, audit security logs, and control agent commission tiers.

Authentication is handled via JWT tokens stored in HttpOnly cookies with role-based route guards verifying permissions before executing controller logic.

2. Property Lifecycle State Machine

Properties move through explicit operational stages: Draft → Pending Review → Active (Published) → Reserved → Rented / Sold → Archived

• When an agent submits a listing, it automatically enters Pending Review and notifies assigned supervisors.
• Supervisors can request listing revisions with structured feedback notes or approve the property to immediately index it for public radial search.

3. Geospatial Property Indexing in MongoDB

• Uses a MongoDB 2D Sphere geospatial index ({ location: "2dsphere" }) on GeoJSON point coordinates ([longitude, latitude]).
• Implemented radial search queries using $nearSphere with $maxDistance parameters, allowing buyers to find properties within a defined radius of work or metro hubs.
• Combined geospatial filters with compound indexes on { status: 1, price: 1, propertyType: 1 } for fast multi-criteria queries.

4. Native WhatsApp Dispatcher via Baileys

Source: @whiskeysockets/baileys and Socket.io

• Managed a native WhatsApp Web socket connection daemon inside Node.js, eliminating third-party per-message SMS gateway fees.
• Session Persistence: Auth credentials and pre-keys are saved atomically to multi-file storage (.runtime/wa-session/).
• Reconnection Resiliency: Implemented exponential backoff reconnection timers and keep-alive ping mechanisms to recover gracefully from mobile network drops.

5. Media Pipeline & Cloudinary Integration

• Property photos undergo server-side validation (file size limits, MIME type verification) before streaming directly to Cloudinary.
• Generates responsive image transformations and optimized thumbnails to reduce mobile data usage during catalog browsing.

Technical Challenges & Trade-offs

• WebSocket State Resiliency: Mobile network disconnects or process restarts required implementing atomic auth file saving and exponential backoff reconnection timers.
• Geospatial & Filter Index Optimization: Combining spatial coordinates with dynamic price and property type filters required tuning compound index keys to avoid in-memory sorting penalties.

Related Resources

• Technologies: NENext.js, NONode.js, EXExpress.js, MOMongoDB, SOSocket.IO
• Direct Contact: Contact Me