Nov 2025 · Rochester, NY
Team Leader and ML Engineer
RIT HACK.COMS
Won first place with UniHear, a platform that turns video and news articles into ASL video. More than 150 students competed.
RIT hosts the largest deaf and hard of hearing college community in the country, so captions being the default felt like the wrong default. We built UniHear over the weekend to turn media into American Sign Language.
We scraped and cleaned hundreds of English to ASL clips into a searchable warehouse, hooked into YouTube transcripts, and wrote a Python service that matched transcript timestamps to the right clips and stitched them together. Google Cloud handled the video work. With a few hours left we added a multilingual layer.
We took first place among more than 150 students. I led the team and owned the machine learning side.