concept project
Aqentix
A concept platform for multi-location inventory: per-store stock signals, reorder workflows, trend visualization, and a data model designed for downstream integrations (POS, supplier APIs).
Problem
What it solves.
Operators juggle spreadsheets across stores. Aqentix unifies the signal surface and lets reorder happen on a schedule, not a panic.
What I built
The specific parts I implemented.
What I owned
My responsibility surface.
Data model
Frontend dashboards
Reorder logic
Architecture
How data moves through the system.
Screenshots
What the system looks like.
Data model
What entities the system tracks.
- id - name - plan
- id - tenant_id - label - address
- id - tenant_id - sku - name - reorder_point
- id - item_id - qty - status - created_at
Failure handling
What happens when something breaks.
Tenant isolation breach
All queries scope by tenant_id; schema enforces RLS-equivalent filter at the application layer.
Reorder rule evaluation timeout
Fall back to manual reorder workflow; log timeout with context for debugging.
Concurrent reorder on same item
Optimistic locking on ReorderEvent creation; second writer receives conflict and retries.
Verification
How I tested or evaluated it.
- ✓Multi-tenant data isolation: tenant A queries never return tenant B items (verified in integration tests).
- ✓Reorder workflow: trigger reorder for an item below threshold and confirm ReorderEvent is created with correct qty.
- ✓Trend aggregation: aggregate sales data across locations and confirm totals match raw input sums.
Next improvements
Known gaps & next moves.
- →Integrations (Square, Shopify).
- →Forecast model behind reorder.
- →Per-role permissioning.