AI Affinity LLC | | Remote (US Timezones) | Full-Time / INTERNS | Citizen/PR only | https://aiaffinity.com

We are a small 3 member experienced dev team looking for a junior data scientist with excellent Python (ML) and SQL chops to work in the healthcare fraud waste and abuse space. Chance to make a large impact while you grow as a data scientist / ML engineer in a stress free environment. Excellent work life balance!

Please email us your resume at: [email protected]

I had worked on an online machine learning method[0] last year that assumes priming effect (similar to the effect you and the GP describe, i.e., the timing effect) and tries to exploit it.

With respect to infinite scrolling/ranking, I have also worked on a simplified model (unfortunately does not capture spacing of likes but captures fatigue based on relevancy) and and algorithm to work with it.[1]

[0] https://arxiv.org/abs/2006.10356: Learning by Repetition: Stochastic Multi-armed Bandits under Priming Effect

[1] https://arxiv.org/abs/1901.07734: Thompson Sampling for a Fatigue-aware Online Recommendation System

Agree with you that figuring out the cultural fit and applicability is super important. I learned a lot firsthand about this while trying to set up a decision support tool (with a supervised learning algorithm underneath) for a few teams in NASCAR back in 2012 [0].

[0]: https://www.liebertpub.com/doi/pdf/10.1089/big.2014.0018