At TerraPay, I deployed Apache Superset dashboards to monitor 200,000 daily transactions and designed a relational mapping schema to standardise error code categorisation.
I also engineered a fault-tolerant Python reconciliation engine that processed 20,000+ transaction records in under 20 seconds. I optimised PostgreSQL pipelines across EU and Asian regions, reducing cross-border query execution time by 30%.
On ScaleRAG, my ongoing distributed document intelligence project, I assembled an asynchronous ingestion pipeline with FastAPI, Celery, and Redis. I also implemented hybrid retrieval with Qdrant, dense embeddings, sparse keyword search, and Cross-Encoder reranking.
I’ve built projects spanning computer vision and smart-grid optimisation, including a PPE monitoring system with a live React dashboard and a Deep Q-Network agent for adaptive load balancing.

