New York City Metropolitan Area
• Sole engineer owning architecture and delivery of an AI platform for nationwide U.S. insurance compliance queries across all 50 states and U.S. territories.
• Delivered a 97% reduction in lookup time (≈5 minutes to <10 seconds) through automated retrieval, reranking, and answer generation.
• Built fault-tolerant Scrapy ingestion pipelines with deduplication and change detection, processing 8k+ unstructured regulatory webpages weekly.
• Engineered an end-to-end RAG pipeline (vLLM + pgvector + CUDA reranker) achieving ~5s retrieval+rereank latency and 87% Recall@10 over a corpus of 70k+ chunks.
• Deployed a containerized production stack behind an Apache reverse proxy, sustaining 99% uptime during pilot.
• Designed Grafana dashboards to analyse user behaviour trends, providing actionable insights to guide product strategy.
• Drove pilot rollout to 40+ users across multiple enterprises; scaling deployment to OCI to support increased usage demand.
• Demoed the product live at SILA Conference 2025 to industry professionals.