Igor Lazarev

• Designed and trained ML model predicting impressions, clicks, and orders from ad bids, improving ad efficiency by ~5% and enabling sellers to plan budgets more accurately.
• Optimized auction bid strategies using statistical modeling and A/B testing, lowering entry barriers for small/medium sellers.
• Integrated ML pipelines with product dashboards (Apache Superset + PySpark), enabling real-time analytics and reducing decision-making latency.
• Automated data aggregation from multiple sources into a unified analytics layer, improving operational efficiency.

Tech stack: Python, SQL, PySpark, XGBoost, Apache Superset, A/B Testing, Model Deployment, ETL

• Designed and analyzed large-scale A/B experiments for ad platforms (Yandex Business, Yandex Direct), producing statistically valid product recommendations.
• Improved ad creative performance through ML-assisted content evaluation, increasing CTR and user engagement.
• Built a conversion value optimization algorithm, increasing average conversion cost by 27% and boosting campaign profitability.
• Enhanced budget allocation algorithms, reducing overspending by 10% and improving ROI.

Tech stack: Python, SQL, Pandas, A/B Testing, Statistical Analysis, Experiment Design

• Developed a churn prediction model for SberPrime subscribers using ML classification, automating retention analytics and improving proactive customer engagement.
• Performed geospatial movement analysis of truckers using transactional data, identifying post-Feb 2022 behavioral shifts in logistics patterns.

Tech stack: Python, SQL, Machine Learning, Classification Models, Geospatial Analysis, Data Visualization

• Built credit risk scoring models with a 20% Gini improvement, enhancing default prediction accuracy.
• Developed customer pre-approval models with +12% Gini gain, reducing decision-making time and increasing approval rates.
• Created real-time process monitoring dashboards (Tableau) for credit approval workflows, enabling faster anomaly detection.

Tech stack: Python, SQL, PySpark, XGBoost, Apache Superset, A/B Testing, Model Deployment, ETL

• Developed ML and deep learning models for business automation.
• Implemented NLP pipelines for customer review analysis, improving insight extraction quality.
• Created cross-platform chatbot systems (Telegram, MSTeams, Skype, Slack, Discord) with deep learning capabilities, automating internal processes.

Tech stack: Python, NLP, Deep Learning, Chatbot Development, API Integration