Pavlo Mysak
Data Scientist
I’m a Data Scientist @ BERA.ai, where I design and estimate Bayesian hierarchical models to measure all things brand equity for Fortune 500 clients. Recent work has combined instrumental-variable identification (inspired by two-stage least squares), Bayesian variable selection grounded in exclusion-restriction logic, and latent variable structures for hard-to-observe channels. I hold an M.S. in Data Science from Boston University and a B.S. in Business Analytics and Economics from SUNY New Paltz.
Previously, I worked on pricing and forecasting @ La Tourangelle, building demand forecasting and price elasticity models.
My interests include machine learning, Bayesian statistics, causal inference, and econometrics.
Technical Skills
Python, SQL, R, Tableau
Currently Working On
- TRACE — A hybrid deep learning architecture for intermittent demand forecasting
- Estimating Latent Volatility Regimes via Bayesian Markov Switching - [notebook]
- homework, probably
Explore My Work
See my projects and research for manuscripts, experiments, and my journal for technical notes and ideas.