Case Studies

Working Systems, Not Slideware

Every project below is a real, running build, using the same architecture we deploy for clients.

Project 1

Odoo WMS Multi-Warehouse Implementation

Challenge
Warehouses on Excel or legacy software can't enforce FEFO, track lots, or trust their stock numbers.
Build
Complete Odoo multi-warehouse implementation: locations and routes, FEFO/FIFO picking, barcode workflows, putaway and reordering rules, running on realistic SKU and order data.
Result
End-to-end warehouse flow (receiving, putaway, picking, delivery) running live and ready to configure for any operation.
OdooPostgreSQLPythonDocker
Project 2

Maritime Vessel Tracking Pipeline

Challenge
Port logistics teams lack real-time visibility on inbound vessels across Rotterdam, Antwerp, and Hamburg.
Build
AIS data ingestion, orchestrated with Airflow, transformed with dbt into PostgreSQL, and served on a live dashboard. Fully containerized (Docker).
Result
ETA visibility per vessel per port, refresh every 10 minutes.
PythonAirflowdbtPostgreSQLDocker
Project 3

Days-of-Supply Alert Engine

Challenge
Distributors discover stock-outs after they happen; overstock silently ties up working capital.
Build
Sales velocity and current stock feed a DOS calculation per SKU, which triggers threshold alerts delivered to WhatsApp/Telegram every morning.
Result
"5 SKUs below 7 days of supply, 3 SKUs overstocked >90 days", in the owner's pocket at 7am.
PythonPostgreSQLdbtTelegram/WhatsApp API
Project 4

Anomaly Detection with AI Explanations

Challenge
Monitoring systems flag anomalies but nobody understands why, so alerts get ignored.
Build
Real-time anomaly detection pipeline + Claude API layer that explains each anomaly in plain language, delivered via Telegram.
Result
Alerts people actually read and act on.
PythonClaude APITelegram Bot API