Alexander has 20+ years experience in the design of DS automated applications for: oil industry (adaptive production control systems), satellite communication (beams control of mobile flat antennas), retail (auto-demand modelling, forecasting, prices optimization and order strategy optimization), credit risk (auto-scorecards development), marketing (recommender systems, auto-strategy optimization of: campaigns, combo-offers, etc.), and other automated solutions for small, medium and large enterprises.Alex has interests in data mining and artificial intelligence, numerical optimization, numerical methods and other math fields in the area of data science. His passion is to design and develop automated DS applications, and all needed tools like algorithms for: training of neural networks, stepwise linear & logistic regression, non-recursive and recursive estimators, decision trees, linear and non-linear optimization methods, including algorithms for constraint and integer optimization, latent semantic analysis, algorithms for data preparation and analysis, linear and extended Kalman filters, algorithms for numerically stable matrix inversion and other algorithms related to engineering and data science.
He teaches (graduate and in-house) courses: Python for DS, Automated ML, DM, Applied AI, Numerical Optimization, Production Control Systems, Adaptive Control Systems and more.His Ph.D. is on automated modelling of the demand in hypermarket chains. These systems are dynamic, time-varying and highly dimensional – with 10^5 outputs and 10^7 potential input variables.His monograph is “Multivariable System Identification”, ISBN 978-954-9489-42-2. The focus is a generalization of the representations of multiple input multiple output (MIMO) regression models, analysis of their specifics and applicability, and related methods and algorithms used within the stages of large MIMO models development.
Now he is an associate prof. at Technical University of Sofia, Chief Scientist & Co-founder at A4E and in-house trainer on DS.