At Conviva, I architected a distributed Locust load-testing system and drove backend refactoring that reduced core-interface P95 latency by 22%.
I built a Python, Pytest, and Playwright automation framework integrated with CircleCI. It runs more than 3,000 regression cases daily and helped raise development-phase defect interception to 95%.
I also designed an enterprise LLM benchmark workflow, with evaluation metrics and datasets supported by automated scoring scripts. I explored LLM-driven browser testing for complex UI interactions.
Earlier, I worked on a photo editor application at 百度(Baidu) and 北京缪客科技有限公司, improving mobile test automation and CI/CD. I also worked on iOS and Android game-application testing at 北京科游迪科技有限公司.

