Developing vision transformer ML model to predict cell types of embryonic data in a spatial context under Dr. Bianca Dumitrascu of the Columbia University Statistics Department

- Creating Excel framework of aggregating market trends of 5 prospective portfolio companies in the embodied AI space using Python scripts to scrape data, which streamlines firm partners’ workflows.- Conducting comprehensive due diligence on startup revenue streams while actively participating in 5 high-stakes pitch meetings per week, driving investment decisions and fostering strategic partnerships.- Reviewing portfolio company investment memos and publishing the team’s insights as a press release on the firm’s website.

Estate Planning Division—Trusts and Wills

Conducted comprehensive analysis of user analytics via Amplitude to precisely target key demographics and inform strategic directions for future product innovation.

- Generated Sunspot Graphical User Interface (GUI): Developed GUI that tracks sunspot movement and flare/CME eruptivity of all sunspots in the Solar Dynamics Observatory catalog.- Trained Computer Vision Model on Sunspot Eruptivity: Trained CNN model on sunspot evolution videos to infer the eruptivity of a given sunspot.

Developed the server and client debug status API, alerting developers to dysfunctional sites and used by Qoom Developers.

2020 Conrad Challenge Finalist in Transforming Education through TechnologyDeveloped and piloted a classroom management application in 50+ classes to intelligently block distracting sites with AI, encourage students to foster a healthier relationship with learning, and reward productivity.Consulted 50+ teachers across Los Altos Christian School, Gunn High School, and Applied Computing Foundation classrooms to understand their concerns about classroom focus and lesson retention.