Quantum Computing – Quantum Algorithms

Developed software packages realizing quantum algorithms, quantum error correction frameworks and transpilation protocols

Retail – Price Recommendation System

Price recommendation system for categories of similar products was built using econometric techniques. The system has discrete choice model in its core and estimates utilities of consumers to understand their response to changes in the prices of products. Price recommendations are based on profit-maximizing framework.

Retail – Emulating Database to Understand Performance of ML-algorithm

Assessment of efficiency of Machine Learning algorithms is sometimes impossible without carrying out real-life experiments affecting current business processes. To avoid systematic changes in client's database during the experiment, software which emulates client's database was built. Results of emulations were used to properly estimate performance of particular model, while initial database was not affected by the experiment.

Banking – Models for Risk Management

Set of Machine Learning models for assessment of clients’ financial performance was built.

Banking – the Kalman Filter to Assess Dynamics of Bank's Portfolio

Specification of the Kalman Filter was built to estimate and predict trend of the large portfolio on the time-horizon of few months.

Internal Research – Regularization of Random Forest

Regularization technique for Random Forest model was introduced and program package implementing this technique was built using Numpy. Regularization works well for data with particular structure. This structure oftentimes can be found in large datasets.