Aaditya Bhilegaonkar

At Arizona State University, I engineered a Milvus RAG pipeline for source-grounded answers across a large research and medical corpus. I also designed an LLM evaluation harness comparing RAG and long-context models across 11 models and 2,300+ QA cases.

Previously at the University of Phoenix, I accelerated analytics-ready data delivery by 30% across 10+ sources through Python and SQL ETL optimization. I restructured S3 lake layouts to improve partition pruning and cut Redshift Spectrum query costs by 70% on weekly 100–200GB datasets.

At Amdocs, I restored billing data integrity for 200M+ telecom subscribers, helping prevent more than $100K per month in revenue leakage. I also led rapid investigation of a 10M+ row Oracle corruption event and improved batch pipeline runtime by 18% on 500–800GB datasets.

I build production-ready data workflows with SQL, Python, dbt, cloud warehouses, orchestration, testing, and documentation. My recent projects combine PySpark, AWS Glue, BigQuery, Snowflake, and agentic AI to automate ingestion, lineage, impact analysis, and data preparation.