- Played a pivotal role in the launch of Bing Chat (now known as Microsoft Copilot) as a core member of the Sydney team, the Responsible AI team, and as one of the first prompt engineers for GPT-4.- Led the development and shipment of the web grounding data pipeline (RAG solution), onboarding 150+ Bing Answers to a latency-optimized inference service, to significantly reduce hallucinations and improve response reliability and freshness in Bing Chat responses.- Selected as one of 20 Microsoft employees to represent and demo the product to journalists and tech enthusiasts at the official Bing Chat product launch. - Led the development, optimization, evaluation, and release of Microsoft's first Natural Language Generation (NLG) model in Production products -- the text prediction feature used across Microsoft Word, Teams, Outlook, and Edge Browser.- Spearheaded Microsoft's Aether Committee's Workstream on Fairness & Inclusiveness in Natural Language models and applications, managing six cross-company, cross-functional teams developing and delivering foundational Responsible AI measurement and mitigation techniques now leveraged across Microsoft.- Established the Microsoft Turing Academic Program (MS-TAP), facilitating AI research collaborations between Microsoft Research, Microsoft Turing, and ten university partners.

- Architected and implemented big data solution to provide users with enhanced query performance, correlation and visualization capabilities, lower latency, and improved security/compliance, while cutting costs by $180,000/year- Identified opportunities to enable machine learning capabilities across the ITSM platform to classify and resolve spam incident tickets (~10% of total monthly incident ticket volume) and to decrease the mean time to respond, mitigate, and resolve incidents

- Designed, piloted, and implemented experiment to determine how the immediacy and expanse of online educational content influences long and short-term retention of computer science concepts- Presented research at the Association for Psychological Sciences (APS) and Psychonomic Society conferences - Advised investigative research on replicating human learning and memory processes in machine learning models

- Taught computer science fundamentals using Python, Web Development and Processing to 15 middle school girls- Encouraged and enabled students to pursue careers in technology by discussing leading women in technical roles, outlining career opportunities, and planning field trips to local tech companies and universities- Maintained an engaging and accessible learning environment through collaborative projects and individual assessments

- Built computer vision model (CNN) to detect defects in engine assembly line with up to 93% accuracy to save top client $3 million/year using Tensorflow’s object detection API, in addition to other open-source tools, running on AWS

- Served as collegiate ambassador to donors, faculty, alumni, and current and prospective students at all UCLA events- Administered oversight to 45-member team and balanced 30 hour/week commitment as three-year member

- Doubled profits by optimizing marketing strategies, tracking growth, and cutting costs by 15%- Organized and executed live auction and international online auction for large-scale fundraising- Increased sponsorship support by 70% and maintained professional relationships with new and previous sponsors

- Launched new brand platform across digital properties and doubled following on social media platforms- Conducted market research on brand culture and target audiences to build customized sales pitches - Compiled competitive analysis to track and predict trends in advertising