Denis Chevkota

• Built the ML function from the ground up: established processes for working with business stakeholders, use-case prioritization, and end-to-end deployment of models into CRM operations. • Formed and led a cross-functional team (Data Science, Data Analytics, MLOps, Project Management): hiring, talent development, goal setting, and performance management. • Implemented multi-channel communication optimization for a customer base of 10M+ users → reduced outreach costs by 12-15%. • Deployed a churn prediction pipeline processing 500K+ events per day → improved customer retention and increased LTV. • Launched an early churn warning system → reduced avoidable customer losses by ~3–5%, protecting a significant portion of annual revenue. • Built an uplift modeling framework for targeted campaigns → increased marketing efficiency by 10-12% and reduced unnecessary communications by 15-18%. • Managed the ML project portfolio: ensured delivery timelines, quality, and measurable business impact; regularly reported and presented results to senior management. • Led strategic and resource planning: developed the AI roadmap, budgeting, and long-term priorities for the ML function.

• Led the ML function in financial forecasting and credit risk modeling (PD, EAD, EL) → improved accuracy of key models by 5-8%, contributing to reduced portfolio risk. • Built and implemented balance forecasting models → increased liquidity management accuracy by 7-10% and reduced deviations in daily cash planning. • Established end-to-end ML lifecycle processes (data → training → deployment → monitoring), ensuring stable adoption of models in core business processes. • Partnered with C-level executives to align the AI roadmap with business strategy → accelerated deployment of ML-driven forecasting tools by 1.5–2.

- Team management. - Project management. Completed projects: • Designed and deployed company-wide churn prediction model → improved retention initiatives and preserved an estimated $3.5M+ in annual revenue. • Built internal NLP/RAG-based AI assistant → reduced information retrieval time by 40% for 20,000+ employees. • Developed computer vision verification pipeline → strengthened security controls and reduced manual review workload by 35%.

- Data collection, preparation, and cleaning. - Developed a model for detecting equipment malfunctions, resulting in a 5% reduction in material costs. - Optimized the forecast of material stock at sites (warehouses), improving resource planning and reducing downtime. - Built reports for monitoring production performance.

- Data collection, preparation, and cleaning. - Development of scoring models for microfinance organizations.

- Data collection, processing, and analysis. - Visualization of analysis results. - Development and creation of data marts. - Optimization of SQL queries.