Izzie Lau
Data Science @ Microsoft in Sunnyvale
About
Hey there!
Education
Computer Science, Cognitive Science, Data Science, and Entrepreneurship & Technology
Work Experience
Implemented small-scale experiments for a self-training architecture developed by Microsoft Research with applications to Bing captioning. Designed and trained 10 different iterations of student-teacher models.
Developed and trained RoBERTa models to predict user engagement based on inputs. Designed metrics, test datasets, and Python scripts to evaluate the model’s performance and effectiveness.
Led product research and client interviews to design an internal web platform for employees. Primary designer of high-fidelity Figma mockups of the application. Implemented and launched a working MVP web application using React and C++. Presented the final product to teams within Bing to market its release.
Pioneered product research to design a novel infrastructure sensor package for autonomous vehicles. Implemented and analyzed OpenCV object detection/classification algorithms on different operating systems and setups.
Projects
Worked closely with the founders to conduct end-to-end product research to improve tigertalk.io, a product for public-speaking empowerment. Created a list of key features/functionalities and designed ideal user flows. Used Figma to design low-fidelity and high-fidelity prototypes, iterated on designs based on client and user feedback.
Product manager of a 10-person team, developed a mobile app that streamlines the diagnosis of craniosynostosis in newborns. Oversaw team using sprints, task boards, weekly standups, and check-ins while building team camaderie. Developed alongside the engineering team in React Native and Typescript.
Implemented new features for sproul.club, a website with 500+ weekly users that brings transparency and marketing to Berkeley’s club ecosystem.