Full-Stack Developer — Build a Multi-Channel WhatsApp CRM & Messaging Platform

We're building a multi-channel customer-messaging and CRM platform — a single real-time inbox that unifies WhatsApp Business (Cloud API), WhatsApp Personal, Instagram DM, and Facebook Messenger with a built-in CRM, a broadcast/campaign engine, and native in-chat payments. Think WATI / AiSensy / Gallabox, but lighter, faster, and tailored for SMBs and creators with paid audiences. We have a clear product spec and need an experienced full-stack engineer who can own this end-to-end (or take ownership of major modules) and — critically — build it to scale and stay maintainable for years, not just ship a demo. Clean architecture, modular code, and sensible data modeling matter as much as features. What you'll build Unified real-time inbox across WhatsApp Cloud API + WhatsApp Personal + Instagram + Messenger, with typing indicators, reactions, read receipts, and per-channel send routing (Socket.IO with WebSocket + polling fallback, sticky-session Nginx). CRM panel inside the inbox — contact profiles linked across channels, tags, notes, per-customer template library, order/timeline view. Campaign / broadcast engine — scheduled, batched dispatch with delivery + click analytics. Native in-WhatsApp payments via Meta Payments Hub (interactive.order_details flow) wired to a PSP (Razorpay / PayU / Cashfree) — customer pays without leaving the chat. Visual chatbot / flow builder (ReactFlow) — drag-and-drop nodes, variables, AI-assisted flow generation. (Phase 2 / nice-to-have.) Webhook ingestion for 4–6 sources with signed-webhook verification and idempotent processing. Hybrid persistence — PostgreSQL (Prisma) for structured CRM/campaigns/billing, MongoDB (Mongoose) for the high-volume chat-message + webhook-event firehose, Redis for sessions/cache. Media + async pipeline — S3 (and GridFS) for media; offload large history-sync batches to SQS + Lambda so the request thread never blocks. Required tech stack Next.js 15 (App Router), TypeScript (strict), React 18 PostgreSQL + Prisma MongoDB + Mongoose Redis (ioredis) Socket.IO (real-time) AWS — S3, SQS, Lambda, EventBridge, CloudWatch WhatsApp Cloud API (Meta Graph) + OAuth flows Tailwind CSS, Shadcn UI Required experience Shipped production full-stack apps on Next.js + Node (please link real work). Comfortable modeling data across both relational (Postgres) and document (Mongo) stores, and knowing when to use which. Real-time systems experience — Socket.IO/WebSocket under load (connection storms, sticky sessions). Third-party API + webhook integration experience; bonus: WhatsApp/Meta, Baileys, or Zalo. Has handled real scaling concerns: connection pooling (e.g. ProxySQL), Redis caching, SQS queues, auto-scaling. Writes clean, modular, well-organized code we can keep extending for years. Nice to have WhatsApp Business API / Baileys / Zalo personal-account bridges Meta Payments Hub / in-chat checkout Shopify / Shopee integration + GDPR webhooks AI integration (Claude / OpenAI) for reply-assist and flow generation Deliverables (milestones) M1 — Foundation: Next.js + TS scaffold, Postgres (Prisma) + Mongo (Mongoose) + Redis wired, auth, base inbox UI. M2 — Channels + real-time: WhatsApp Cloud API + one more channel into a live Socket.IO inbox with webhooks. M3 — CRM + campaigns: contact/CRM panel, tags/notes/templates, broadcast engine with analytics. M4 — Payments + hardening: in-WhatsApp payments (Meta Payments Hub + PSP), S3/SQS media + async pipeline, load testing.