Supriya Pasupuleti

Engineered an innovative tool to detect missing transactions during clinical data conversion, reducing 80% of manual work involved. Collaborated on automating migration of 1000+ Python programs across studies, connecting to patient insights database. Integrated GPT-3.5 Turbo model to generate Python codes with 70% accuracy, assessing data quality using few-shot prompting. Built tools automating data management, reducing communication gaps between developers and testers. Acquired proficiency in Git Actions, automating post-push tasks, code review, and validation for 100+ codes with comprehensive logs. Coordinated development of robust APIs for search functionality in 4 formats: exact, starts with, ends with, and contains search. Customized AWS Lambda function to automate PostgreSQL database updates, reducing 20 hours of manual work weekly. Achieved up to 70% optimization of advanced PostgreSQL queries, boosting data retrieval efficiency and performance.

Enhanced Python Flask APIs to achieve 100% efficiency in extracting knowledge from PDFs, text, and web URLs. Deployed locally built Python scripts as Cloud functions and Containers for version control and 50% faster execution time. Configured MongoDB for efficient storage and retrieval of diverse knowledge extracted from text, web pages, and PDFs. Performed conversion of MongoDB-stored data into Neo4j graph database, enhancing data representation and analysis capabilities. Implemented Python and Shell scripting codes for seamless movement of over 200 files per hour from GCS to BigQuery. Contributed to Python scripts for converting nested XML and JSON data, up to 6 degrees, into CSV format. Involved in creating and updating PostgreSQL metadata tables to ensure consistency and eliminate redundancies.