Can I vibe code MoEngage?

Build a self-hosted customer engagement backend microservice using Node.js, TypeScript, PostgreSQL, and BullMQ.

1. DATA MODEL:
- Users: id, email, phone, attributes (JSONB), opt_in_email (boolean), opt_in_sms (boolean), created_at.
- Events: id, user_id, event_name, properties (JSONB), timestamp.
- Segments: id, name, filter_rules (JSONB logic for attribute and event criteria).
- Campaigns: id, name, channel ('email'|'sms'), trigger_type ('event'|'scheduled'), segment_id, template_body, status.
- DeliveryLogs: id, campaign_id, user_id, channel, status ('queued'|'sent'|'failed'), provider_response (JSONB), sent_at.

2. CORE FUNCTIONALITY:
- Provide an API endpoint POST /api/v1/events to ingest user events, update user attributes, and append to the Events table.
- Implement a real-time trigger evaluator: when an event matches a active Campaign's trigger criteria (e.g., event_name === 'cart_abandoned'), queue a message job in BullMQ.
- Implement a worker process that consumes BullMQ jobs, checks user opt-in flags, renders liquid templates with event properties, and dispatches via external APIs (SendGrid for email, Twilio for SMS).
- Handle webhook endpoints POST /api/v1/webhooks/delivery to update DeliveryLogs status on bounces or delivery confirmations.

3. FAILURE MODES & RESILIENCY:
- Implement exponential backoff retries for third-party API rate limits (HTTP 429).
- Idempotency key evaluation using Redis to prevent duplicate messages sent to the same user within a defined window.
- Hard failure handling: automatically mark opt_in_email = false when receiving a hard bounce webhook.

4. OUT OF SCOPE:
- Native mobile push SDKs (APNs/FCM handling).
- Drag-and-drop visual flowchart UI.
- Predictive ML models for send-time optimization.

$ each button prefixes agent-specific run instructions · build your own product, never copy proprietary code, trademarks or designs