Working on Generative AI solutions using Python, LLMs (Llama), LangChain, LangGraph and vector databases like Milvus. Focused on building AI agents, semantic search capabilities, and natural language-based querying, collaborating with a team of interns and mentors. Documenting learnings and innovations throughout the process.

Implemented an ETL pipeline utilizing AWS Step Function and Glue to ingest structured file-based data into S3. This automated solution, deployed through CI/CD pipelines, replaced the previous manual process, resulting in approximately 60% savings in operational costs.Maintained data integrity through automated data quality checks and validation scripts within the ETL process, reducing errors and inconsistencies.Leveraged AWS Glue Data Catalog to centrally manage and maintain metadata for datasets ingested through the ETL pipeline, enhancing data discoverability and consistency, and integrated it with AWS Glue jobs and Athena to streamline query execution and improve data analysis efficiency.Utilized AWS Identity and Access Management (IAM) to enforce strict access controls and permissions, safeguarding data against unauthorized access.Implemented monitoring and alerting mechanisms using AWS CloudWatch and AWS SNS to proactively identify and resolve data discrepancies and pipeline issues. Developed and maintained comprehensive documentation of data workflows, ETL processes, and data governance policies to support ongoing maintenance and future enhancements.

Designed and implemented an automated data pipeline using Python to ingest unstructured file data with metadata from an API endpoint into an S3 data lake. This solution enabled the data science team to build ML models that generated critical insights into clinical trial progress. By reducing manual data processing time and costs, the automated pipeline facilitated faster decision-making and improved operational efficiency, ultimately resulting in significant cost savings and increased revenue potential.Developed reusable components using TypeScript and built responsive dashboards with HTML5 and CSS, ensuring seamless navigation, cross-browser compatibility, and integration of key KPIs for laboratory testing and dosage efficacy.Applied statistical techniques to identify factors impacting product efficacy and sales performance contributing to a 10% increase in sales and reduced stockouts. Developed and optimized complex SQL queries using joins, subqueries, and window functions to extract actionable insights and support data-driven decision-making. With Denodo, successfully combined data sources such as Oracle, MS-SQL, Amazon Redshift, and Web services securely serving information to consuming applications such as Tableau and Excel. Effectively communicated business value to stakeholders, aligning project goals with technical implementations to drive informed decision-making and optimize project outcomes.