Professional Summary
Hands-on GenAI Engineer with 9+ years of experience building production-grade RAG systems in Azure. Designed scalable ingestion pipelines for 10K+ enterprise PDFs, implemented metadata-driven access control using Unity Catalog and Azure AD, and deployed GPT-powered APIs serving 2,000+ daily users. Strong expertise in Azure OpenAI, Databricks, vector search, and secure API architecture.
Key Production Experience
Enterprise RAG Platform Healthcare Client
- Designed ingestion pipeline for 12,000+ clinical documents using Azure Document Intelligence.
- Implemented structure-aware chunking with table handling.
- Built embedding pipeline using Azure OpenAI and Azure AI Search.
- Implemented department-level filtering via metadata + Azure AD claims.
- Reduced hallucination rate by 32% using retrieval evaluation and prompt enforcement.
- Deployed FastAPI service with Azure API Management and OAuth2 authentication.
- Processed 5,000 queries/day with <2s latency.
- Implemented token cost monitoring and caching strategy reducing cost by 28%.
Databricks-Based RAG Migration Financial Client
- Built Bronze/Silver/Gold Delta architecture for document lifecycle.
- Used Unity Catalog for row-level security.
- Managed notebook CI/CD using GitLab.
- Implemented incremental embedding for versioned documents.
Technical Skills
- Azure OpenAI, Azure AI Search, Azure Functions
- Databricks, Unity Catalog, PySpark
- FastAPI, .NET API, OAuth2, Azure AD
- LangChain (used sparingly in production)
- Redis vector search
- MLflow, CI/CD (GitHub Actions)
- Monitoring via Azure Monitor & Prometheus
Metrics Section (Very Important)
Strong CVs include:
- Number of documents handled
- Daily query volume
- Latency improvements
- Cost reduction %
- Evaluation score improvements
- Uptime %
Weak CVs avoid metrics.
