ENGINEERING AT MERXIQ
Engineered for scale.
Engineered for scale.
Built for trust.
A cloud-native stack — modern frontend, a battle-tested backend, and a multi-model AI orchestration layer — running on infrastructure designed for multi-tenant SaaS from day one.
POST /v1/ai/insights // tenant-scoped, routed through the AI orchestration layer { "tenant": "school_482", "signal": "attendance_risk", "model": "claude" } → 200 OK
The Stack
One stack, every layer
Each layer picked for a reason — fast to ship on, boring where it counts, and able to carry real production load.
Frontend
Next.js
React
Tailwind CSS
Backend
Laravel
Database
PostgreSQL
Redis
AI Layer
Azure OpenAI
Claude
NVIDIA NIM
AWS Bedrock
Cloud & Infrastructure
Azure
AWS
Docker
Kubernetes
System Architecture
How a request moves through MerxIQ
From a browser to a tenant-scoped response, every layer runs independently scalable and independently secured.
Client
Browser / Mobile Web
Application
Next.js + React Frontend
Laravel API
Data
PostgreSQL
Redis Cache
AI Orchestration
Azure OpenAI
Claude
NVIDIA NIM
AWS Bedrock
Running on Docker + Kubernetes, across Azure and AWS
Multi-Tenant by Design
One platform. Every school isolated.
MerxIQ runs as a single shared platform across schools, with strict logical isolation at every layer — so one tenant's data, load, or incident never touches another's.
- Every row is scoped to a tenant ID, enforced at the database layer — not just in application code
- Each tenant gets its own Redis namespace for sessions and cache, so one school's traffic can't evict another's
- Shared compute, isolated blast radius — a problem in one tenant's workload doesn't degrade another's
School A
School B
School C
MerxIQ Platform — Shared Application Layer
Tenant-scoped rows
PostgreSQL
PostgreSQL
Isolated namespace
Redis
Redis
Per-tenant audit log
Postgres + logs
Postgres + logs
Enterprise-Grade
Built to scale. Built to protect.
Scalability
- Stateless application layer, horizontally auto-scaled with Kubernetes
- Redis-backed caching to absorb read-heavy traffic without hitting the database
- AI workloads run asynchronously, so a slow model call never blocks the request path
- Containerized services deploy independently across Azure and AWS
Security
- Encryption in transit (TLS) and at rest across every data store
- Role-based access control on every API endpoint and admin action
- Tenant data isolation enforced at the database layer, not just in app logic
- Audit-logged admin and AI-assisted actions for full traceability
Want the deeper technical walkthrough?
Happy to go layer by layer with your engineering team. Contact us