MIT App Inventor's Conversational AI Team

Helped develop Aptly, a tool that converts English descriptions into App Inventor apps.

Implemented App Inventor features such as comments and procedures in the Aptly programming language (Python).

Working on functionality to allow users to describe desired edits to their app in natural language and have those changes be automatically implemented (Python, JavaScript, Java).

Led recitations to guide MIT students through learning new technologies and review key concepts from lecture (TypeScript, Vue.js, Node.js, HTML/CSS)

Prepared recitation lesson plans (outlines, slide decks, and interactive activities) and preparatory assignments for students

Provided individualized assistance to students in office hours and provided detailed feedback on student assignment submissions

Conserved network usage for both Google and quota-constrained buyers by developing a new control in Google Display Ads Real-Time Bidding for advertisers to only receive bid requests for ad impressions they would be interested in (C++).

Reduced compute resource usage and susceptibility to bugs by refactoring currency-related code to eliminate unnecessary currency type conversions and improve code maintainability.

Clarified the semantics of the Google RTB protocol by populating currency-related fields to reduce advertiser confusion regarding the currency of money amounts being sent on bid requests.

Further aligned Google’s bid requests with OpenRTB industry standards by populating the available currencies on bid requests.