Tommy Mulc
Former Data Scientist and ML Researcher at Google and Expedia. Here are some things I’ve worked on.
- Expedia’s Performance Marketing bidding algorithms. Large-scale ML with a large impact on bottom-line performance and long-standing legacy.
- Open-source implementation of Tacotron. One of the first reproduced works of the popular text-to-speech model.
- Google Search Trends opportunities. Helped clients uncover Search keywords that were trending and relevant.
- Media Mix Modeling with Neural Networks. An alternative to traditional MMMs that allow users to measure the creative impact of their ads.
GitHub
Hackathons
My favorite hackathon project was a device that found your face and shot candy into your mouth. You can checkout a video here.
Papers
- NNN (alternative to MMMs)
- Q-Match (self-supervised learning improvements for tabular data)
- Compressing Search with Language Models
- CSS10 (text-to-speech datasets)
- Undergraduate math thesis: varianceclusteringsmulc
- Ebola eradication (MCM 2015): mcm_2015