Revanth Y.

Experience

1. Ai Ml Engineer

   CVS Health

   Oct 2025 — Present

   Built an end-to-end document ingestion and extraction pipeline using Document AI, capable of parsing complex invoice formats and accurately extracting header attributes, line items, vendor information, and purchase order references at scale. Implemented advanced data normalization and validation mechanisms, incorporating schema enforcement and confidence-based filtering to improve extraction accuracy and ensure high-quality structured outputs. Engineered a comprehensive exception detection framework to identify discrepancies such as duplicate invoices, PO mismatches, pricing inconsistencies, and threshold violations based on business rules and historical patterns. Developed a hybrid decisioning system integrating rule-based validation with machine learning models using BigQuery ML for anomaly detection and probabilistic risk scoring of invoice records. Leveraged BigQuery Vector Search to implement similarity-based retrieval of historical exception cases, improving contextual understanding and enabling more accurate resolution recommendations. Designed and implemented a retrieval-augmented generation (RAG) pipeline, enriching LLM prompts with structured invoice data, rule evaluation outputs, vendor context, and historical exception datasets to enhance reasoning capabilities.