At ServiceNow, I co-developed an internal Python framework with a Staff Engineer to automate multi-table synthetic data generation for Generative AI golden datasets. I enhanced it with AI-assisted Windsurf workflows, reducing dataset generation time from several hours to minutes.
I generated 30+ golden datasets to evaluate Generative AI features and agents, and analyzed evaluations to identify failures across prompts, datasets, judge models, and platform behavior. I also performed root cause analysis and partnered with engineering teams to improve evaluation reliability.
As a Software Engineer at ServiceNow, I developed Selenium and Playwright automation suites that helped improve regression test automation coverage from <50% to >90% as part of the Greenlight Initiative.
Previously, at Tata Consultancy Services, I engineered Python, Flask, and PostgreSQL backend services for an enterprise pharmacovigilance platform supporting adverse event reporting. I implemented RabbitMQ and Celery asynchronous workflows and optimized PostgreSQL queries and indexing strategies, improving API response time by 30%.
