Distributed Event Bus & CDC Layer
Implemented Change Data Capture (CDC) over legacy transactional databases, streaming inventory state mutations into a centralized Apache Kafka message bus with sub-100ms propagation.
Case study · Supply Chain & Distribution
The challenge
Operating across multi-country distribution hubs, regional teams relied on asynchronous batch exports from legacy ERPs combined with local spreadsheets. Order exceptions required manual investigation across five separate systems, creating multi-day fulfillment delays during peak demand.
The approach
Implemented Change Data Capture (CDC) over legacy transactional databases, streaming inventory state mutations into a centralized Apache Kafka message bus with sub-100ms propagation.
Replaced unstructured email and spreadsheet escalations with a state-machine-driven dispatch engine that assigns, audits, and auto-resolves inventory discrepancies.
Built a unified, low-latency web cockpit providing floor managers and executive leadership with real-time throughput metrics, bottlenecks, and automated SLA alerts.
Reported outcomes
Results are described from the engagement context. Client details are anonymized where confidentiality requires it.
74%
Reduced average investigation time from 18 hours to under 4.5 hours.
99.98%
Zero data divergence between warehouse floor actions and corporate inventory records.
4.2x
Handled record seasonal volume spikes with zero system throttling or dropped orders.
Tools used in this engagement
Next.js · TypeScript · Node.js · Kafka · PostgreSQL · Redis · Docker