Worked with various clients to deliver data-driven solutions across multiple industries.Client: Caremetx, LLC | Dec 2023 – Jun 2024Led the optimization of SQL Server procedures, enhancing query performance by 80%.Integrated Confluent Kafka with AWS Lambda to streamline data flow, reducing data processing times and meeting strict project deadlines.Collaborated with cross-functional teams to integrate new APIs into legacy systems, improving system interoperability.Led the development of an AWS-based document transfer pipeline, significantly improving data handling efficiency.Designed and launched a real-time API endpoint for non-prescription updates, ensuring 100% data accuracy across SQL Server tables.Client: Bayer Customer360 | Feb 2022 – Nov 2023Orchestrated the migration of repositories to GitHub Enterprise Cloud, integrating AWS IAM and GitHub Actions to enhance CI/CD processes.Developed robust features for a GraphQL API, enabling real-time customer data delivery.Resolved critical bugs in a PostgreSQL environment, ensuring seamless data transactions and system reliability.

Built and maintained insurance data pipeline orchestration

Led training of annotation specialists and offshore contractors using homegrown software Workbench and Excel for capturing and controlling the quality of P&C insurance data in an unstructured image data to structured, normalized JSON and CSV files. Pulled data from Goolgle BigQuery data warehouse and utilized Python and Jupyter Notebook to classify Insurance Policy page types.

Data Managemento Developed standard procedure for data intake procedure to transfer, pre-process, load, and distribute data tointernal and external researchers.o Designed meta-data schema for large database to efficiently store meta-data of data tables and technical notes toeach dataset.o Built Python scripts and T-SQL stored procedures to load at most 27.9 million rows of education data and its metadata into Michigan Education Data Center (MEDC) Database successfully.o Optimized data tables and meta-data schema by utilizing relational structure of database.Automationo Created Python script for generating a report of group membership and directory information and for controllinggroup access using multiple APIs provided by University of Michigan IT Services.o Automated process of decrypting zip and 7z compressed AES-256 encrypted data files into network storage andlogging details into shared spreadsheet; 60+ GB of data uncompressed without manual process.

Undergraduate Research Assistant at Public Design WorkshopData-driven videography and motion graphics