Sanin Adnan

About Sanin Adnan

Looking for a position that utilizes my strong technical background and collaborative skills to build applications in a team environment.

Experience

1. Principal Engineer

   CVS Health

   Mar 2025 — Present

   Architected an end-to-end RAG system integrated with Slack, enabling natural language querying across disparate internal data sources using a PostgreSQL vector database- Engineered a high-fidelity ingestion pipeline using Docling and Haystack to transform unstructured data into Markdown, implementing optimized chunking and multi-model embedding strategies- Optimized retrieval accuracy by implementing a hybrid search (BM25 + Vector) architecture, utilizing LLM-driven pre-filtering to restrict search space and Cross-Encoders for high-precision reranking.• Established LLM evaluation suite using DeepEval (RAGAS) and Arize Phoenix to monitor faithfulness and relevance, using golden datasets and user feedback loops for iterative tuning.• Developed Agentic RPA workflows using Azure Foundry OpenAI and Browser-Use to automate complex human tasks, leveraging Python Celery for asynchronous execution and distributed task management.• Managed full-cycle deployment of AI services using Docker, Kubernetes, and ArgoCD, ensuring scalable infrastructure for agentic monitoring and observability.

Education

• University of Virginia

  Bachelor's degree, Computer Science and Cognitive Psychology

  2002 — 2006

• George Mason University

  Masters, Information Security

  2010 — 2012