At Publicis Sapient, I established a structured way to validate AI-generated output when no test method existed. I designed and maintained Playwright end-to-end test suites across UI workflows, REST APIs, and underlying data, and contributed to a 15% reduction in production defects.
For Backlog AI, an AI-powered Jira ticket generation tool, I was the sole quality owner, carrying test strategy, automation, and defect resolution end to end. I agreed acceptance criteria with product managers and developers before build started, then wrote and maintained Cypress regression scripts.
At Sky Limited, I built the Karate API testing framework from zero, taking coverage from 0% to 80%; the team later adopted it as its standard. Earlier at Infosys Limited, I built a self-service subscriber dashboard that cut user query tickets by 40% and optimized batch jobs and real-time address cleansing for a 10% performance gain.
