Postgraduate student at KPI IASA. MS in Statistics at Taras Schevchenko KNU. 4 years experience in Data Science field. Worked in financial, consulting and IT spheres. Have experience as python backend developer and PM. Scientific interests lie in a field of creating real-time scoring models under the big data concept and their efficiency valuation.
Topic: AdTech Predictive Bidding Model
Short Description: The presentation concerned to the technology used for building predictive bidding model for estimating probability of user action as response on the seen ad. Predictive model is based on historical auctions data and results of their conversions collected by QupleTech. The core statistical engine of the model is built on logistic regression trained via stochastic gradient descent optimization principle with adaptive learning rate. An original approach was used to valuate the model efficiency.