kkaliuzh

I built a Python research pipeline for financial market and order-book data, covering automated collection, quality controls, feature engineering, time-series modelling and realistic backtesting. I compared machine-learning models with simple baselines and applied walk-forward validation and leakage controls. This strengthened my ability to structure ambiguous problems, improve processes, document assumptions and distinguish robust evidence from attractive but fragile results.
 My physics education trained me to reason quantitatively, while my current work in special education has strengthened my resilience, communication and responsibility within a team