San Francisco, California, United States
• Led the web client implementation work for new Safety Mode feature, an AI-driven content moderation tool that autoblocks abusive content (using Twitter’s pBlock, pTox, and profanityFilter ML models) helping users to feel safe on the platform, resulting in increased retention rates, UAM and mDAU.
• Implemented UI for users to interact with the tool and to provide feedback on problematic model outputs (false positives/negatives), which helped us to enhance model’s performance and capabilities. Worked primarily with Javascript, React, Redux, GraphQL, Relay, and APIs.
• Collaborated with the product, design, research, ML, backend and mobile client engineers, authored/co-authored technical design documentation.
• Participated in on-call rotation providing operational support for production systems and debugging issues; conducted postmortems.
• Implemented A/B, unit, and integration testing; created detailed test plan documents and coordinated E2E testing bug bash sessions with QA testers.