His responsibilities include algorithm development, machine learning model training, model testing and data wrangling.
Mike combines his expertise in math, residential real estate, data wrangling and machine learning to build and refine WA’s products.
Mike was instrumental in the design and development of Valpro WA’s leading AVM. He designed several valuation algorithms for Valpro. These algorithms leverage a wide variety of techniques including K-nearest neighbor, gradient boosting, classification trees and regression.
Mike made major contributions to the development of condition adjusted ValPro (AVV). WA uses condition scores from photos analysis, market data and a WA trained Neural Networks to produce AVV.
He has also done data analysis and model design for bespoke projects related to distressed discounts, rental properties, and loan portfolio matching.
Mike has a BS in mathematics from MIT and a Masters in mathematics from the University of Chicago.