About

Matt Johnson is a Senior Machine Learning Engineer at Spotify, where he has continued to work on their recommender system and has transitioned to marketing optimization work. He has experience running randomized controlled trials to build heterogeneous treatment effect models that optimize marketing programs. Matt has also researched the long-term effects of these models. Prior to his work at Spotify, Matt was a Staff Machine Learning Engineer at ibotta, Inc., where he led the research, design, and implementation of all fraud models. He has experience with languages such as Python, SQL, and R, as well as tools like Spark, Airflow, SageMaker, and sklearn. Before joining ibotta, Matt was a Data Scientist at Haystax Technology, where he worked at the intersection of data science and data engineering with a domain focus in insider threat. He wrote productized software to process data for a Bayesian network that models insider threat and implemented machine learning techniques that yield outputs that can be applied as inputs to a Bayesian network in a scalable manner. Matt has experience with languages such as Python, Javascript, and Java, as well as tools like sklearn, netica, nltk, and d3.js. Matt holds a Master of Science degree in Applied Physics and a Bachelor of Science degree in Physics, both from George Mason University. He also holds two Associate of Applied Science degrees in Information Systems Technology, one in Network Engineering and the other in Programming/Systems Analysis, from Lord Fairfax Community College. Matt completed his high school education at the American School of Correspondence. In addition to his education and work experience, Matt has a strong technical skill set that includes ML, SQL, Scikit, Python, Sagemaker, Spark, Apache Airflow, Pytorch, Naive Bayes, Apache storm, Apache Hive, NLP, MongoDB, and more. He has almost 8 years of relevant experience in the field.