Salimzhan Shaimerdenov

ML Engineer

ALT University

Built forecasting, classification, computer-vision, and NLP models; reported about 15% lower forecast error and an F1-score increase from 0.78 to 0.86.

• Developed and trained ML models for classification, regression, forecasting, and prediction using Python, scikit-learn, and TensorFlow, supporting more accurate business decisions • Built forecasting pipelines with Prophet, ARIMA, LightGBM, and CatBoost, incorporating lag features, rolling statistics, and time-series validation, reducing forecast error (MAPE) by approximately 15% • Performed data preprocessing, feature engineering, exploratory data analysis, and outlier detection on structured and unstructured datasets, resulting in cleaner data for reliable model training

ML Engineer

National Academy of Science

Conducted statistical analysis and feature selection, reducing the feature set by about 20% without loss of accuracy.

• Conducted statistical analysis and hypothesis testing in Python (pandas, SciPy) on Linux, identifying data-quality issues before they reached modeling • Validated statistical significance of key variables using p-values (SciPy), confirming which features meaningfully improved model performance • Analyzed correlations and feature impact using SQL and Python (pandas, seaborn), reducing the feature set by about 20% without loss of accuracy • Delivered clean datasets, visualizations, and scikit-learn models on Linux, improving prediction reliability for business tasks

Bachelor's Degree, Robotics

Satbayev University