Jack Kearney

Background

About Jack Kearney

I began my career as a theoretical physicist; when you love solving challenging problems, it's a hard field to avoid. My physics education gave me a strong foundation in linear algebra, modeling, scientific writing and caffeine management that continues to serve me well. Meanwhile, my time as a postdoc allowed me to hone my skills as an independent researcher and, as a bonus, to travel broadly!After my postdoc, attracted by the revolution happening in machine learning and natural language processing (NLP), I joined the Alexa AI group at Amazon. In my "industry PhD," I learned a huge amount, became an infinitely better software developer, and contributed to Alexa's natural language understanding (NLU) through innovations in active learning, informational retrieval, data weighting and model privacy.Have you ever asked Alexa for music similar to a particular artist or song? Well, I am proud to have played a significant role in that; before me, asking for music "like Elton John" got you "Elton John's greatest hits."After Amazon, I joined Jump Trading, where I explored more statistical (as opposed to product-driven) applications of machine learning. The nimbleness of a smaller company, coupled with my expertise and seniority, allowed me to broaden my research portfolio. I was also worked with all levels of the company, from interns to senior officers, and from data pipelines to model deployment.Using my versatile toolkit and diverse perspective to tackle a wide range of complex, ambiguous problems make me happy. One of my colleagues once said I "consistently generate new, course-altering ideas," and I couldn't be prouder (see Sam Kater's recommendation below). My research expertise includes natural language, model safety/privacy, recommendation systems, reinforcement learning, learning-to-rank, positive-unlabeled learning, large language models, multiple hypothesis testing, change point/anomaly detectionand, of course, particle physics.