Working on the Roblox VR experience. One of a handful of engineers who built the Roblox app for Meta Quest. Now working on sound and haptics in the game engine.

I worked in Prof. Roman Feiman's lab on a joint project between the Brown Language and Thought Lab, and Prof. Ellie Pavlick's NLP lab. We studied how children learn negation terms such as 'not', 'no', and 'nt' with the objective of training deep learning models in a similar manner.

Worked on Sana Voice, a speech pronunciation tool that helps non-Native English speakers learn English. I designed the phoneme-level remediations which help users adjust their articulations to make the target English sound. Scalable design based on IPA's phoneme traits allows for dynamic feedback based on user's native language. Written in Python.

I added a UI for these remediations on the API's web demo. Written in React.

I also created a 'Speech Coach' web app for our API in which user repeats after the computer pronunciation of a phrase. App compiles a list of frequent errors, practices them, and provides user with breakdown of their most common speech errors. Written in React.

• In Python, created HealthLit, a search engine that uses Natural Language Processing to intelligently organize medical literature from clinicaltrials.gov.

• Used “Semantic Role Labeling” & medical taxonomy SNOMED to solve design challenges.

• Findings later published by Association of Computational Linguistics.

• Worked with chair of psychology department, Daniel Swingley, to research how infants learn language.

• Led “language-guided looking experiments” on subjects and designed visual and auditory test queues.

• Used R to visualize and analyze newly-acquired data of 70+ participants across 4 studies.