Unified Data Platform for Supply Chain & Product Information Management
-
Designed and evolved a unified data platform for supply chain and product information management, consolidating data from multiple retail brands and countries into a consistent domain model. The platform integrates product catalogs, point-of-sale inventory, marketplace information, and enterprise PIM systems such as Akeneo and Mirakl through a unified API layer for downstream applications and services.
Key responsibilities and achievements
-
Designed and implemented a distributed event-driven with Kappa architecture, enabling continuous synchronization and transformation of business data across heterogeneous source systems.
-
Built Kafka-based streaming pipelines with Akka services to ingest, process, and transform product, inventory, and commercial data into optimized read models.
-
Designed a unified API layer exposing consistent access to business data independently of underlying source systems, simplifying integration for downstream applications.
- Optimized platform scalability, reliability, and performance to support daily synchronization of all points of sale within a strict four-hour overnight processing window, including:
- Performance optimization and endurance testing
- Continuous monitoring and operational diagnostics
- Metrics collection and observability through the ELK stack
- Production reliability improvements under sustained high throughput
- Managed and optimized distributed data stores operated by the engineering team, including Cassandra and Elasticsearch, with focus on:
- Database tuning and capacity planning
- Performance optimization and reliability improvements
- Collaboration with infrastructure teams on production operations
- Evaluation of MongoDB for specific data access patterns
-
Contributed to technical leadership within a cross-functional squad, collaborating with software engineers, infrastructure specialists, and product owners on architecture decisions, technical trade-offs, and platform evolution.
- Established modular architecture principles allowing ingestion, synchronization, storage, and data exposure capabilities to evolve independently while sharing common data models and operational standards.