Backend Developer
Commerzbank - Frankfurt, Germany
- Reduced average RAG query latency by approximately 70% through application-level caching and optimizations to Azure AI Search.
- Built FastAPI endpoints for dynamically reprovisioning search indexes, enabling teams to experiment with chunking, indexing, and retrieval parameters.
- Designed the integration of Confluence as a data source for the platform, defining the ETL flow from enterprise content into the RAG pipeline.
- Improved production RAG services through metadata handling, API enhancements, bug fixes, and retrieval pipeline improvements.
FastAPI Azure AI Search RAG Confluence