Director, Machine Learning
Matyas Tamas is Director of Machine Learning at Manifold, where he works closely with our customers to translate their needs into machine learning and data engineering systems—from exploratory data analysis, to writing production code. Throughout his career, Matyas has worked with diverse data sets and problem-solving techniques—including experiment design, NLP (natural language processing), behavioral analytics and modeling, economic modeling, time series forecasting, and marketing and attribution modeling. Recent product successes include a text classification machine learning model for an analytics software company, a lift prediction model prototype for a health IT company, and a big data production model pipeline for a public media company.
Prior to Manifold, Matyas co-founded Ontiq, where he developed a data pipeline to cleanse pharmacy data for analysis. He designed and developed machine learning anomaly detection models to analyze patient behavior, to research how different communication patterns with patients might impact patient adherence. Before Ontiq, Matyas built and led the data science team for Uber's Supply Team, a large sub-organization (100+ members) focused on driver growth, engagement, and retention. While there, he also advised on economics-related products—including the dynamic pricing algorithm, UberPOOL, supply positioning, and marketplace balance. Prior to his role at Uber, Matyas created the data science team at Quora. There, he worked on building Quora's data infrastructure, growth and product analytics, experimentation design, email and notification optimization, and using NLP to improve the user experience.
Matyas has a BA in Physics from Caltech.