During my professional career I’ve worked on many different data science projects, but I’ve collected the most extensive experience in implementing novel machine learning solutions in the financial industry. The disruptive nature of Fintech pushed me to coming up with innovative solutions for doing classic credit scoring. For the past couple of years, I have mainly worked on developing highly predictive credit scoring models based on non-traditional data sources and bank account transactional data.
The models I have developed aimed at automating loan processing as much as possible, while still providing high predictive power. They were proved to meet their ambitious goals both internally, as well as with external partners and institutions.
My main areas of financial services modelling expertise are:
- aggregating raw bank account transactions into higher level attributes
- classifying bank account transactions with neural networks
- applying unsupervised ML techniques to bank account transaction descriptions
- modelling and forecasting bank account EDB (Ending-Day-Balance) time-series
- performing EDB time-series clustering and classification
All of the above result in the creation of powerful and highly predictive scoring attributes based solely on bank account transactional data.
In case you find my profile interesting, please reach out to me on Linkedin or write me an email at: firstname.lastname@example.org.