Developed predictive maintenance models achieving 99% accuracy with Random Forest; built a Python-based data pipeline and deployed the optimized model via Streamlit for real-time machine downtime prediction, reducing unplanned maintenance and improving efficiency. Designed an AI-powered chatbot using OpenAI GPT for FDA drug labeling information retrieval; implemented PDF parsing and semantic validation, achieving high contextual accuracy and minimizing hallucinations.

Senior Associate, Analytics at FactSet

March 1, 2021 - March 1, 2023

Built scalable ETL pipelines in Databricks using Medallion architecture, enabling raw-to-curated data processing. Applied SCD2, schema evolution, partitioning & bucketing to manage data versioning and performance. Optimized PySpark jobs (broadcasting, custom partitions, caching) to reduce skew and improve reliability. Automated catch-up/late-arriving data loads and enabled real-time exports for reporting needs. Managed terabytes of daily ingestion with trigger-based and scheduled workflows. Streamlined deployments via Azure DevOps CI/CD, cutting post-release issues by 15%.

Software Engineer at IBM

June 1, 2018 - March 1, 2021

Enhanced data quality and governance by implementing the Great Expectations framework and telemetry validation checks, reducing junk data to <5% across releases. Designed and configured new data attributes, metrics, and events for telemetry pipelines, ensuring accuracy in product analytics and monitoring. Automated volume monitoring framework with thresholds and anomaly detection to proactively flag data drifts. Improved release reliability through CI/CD integration, testing automation, and triaging issues, cutting validation time by 20%.