Denis Miroshnichenko

Responsibilities: - Designed and maintained ETL pipelines (Spark/Scala, Hadoop → PostgreSQL) and automated workflows via Airflow and Kafka for real-time ingestion and ML deployment. - Built dashboards (Grafana, Kibana) and mentored engineers. Achievements: - Optimized ETL runtime by 50%; - Migrated scheduling to Airflow in 2 months; - Cut ML deployment from 6 to 2.5 months; - Improved Spark speed 1.5× across 150+ runs; - Reduced data latency from weeks to hours.

Responsibilities: - Built and optimized data pipelines (Oracle PL/SQL, PySpark, Python) for antifraud, underwriting, and credit risk. - Developed ETL workflows, DWH marts, and BI dashboards for analytics teams. Achievements: - Cut fraudulent approvals by 27% via decision-tree triggers; - Improved scoring accuracy by 4% and speed by 40% with CatBoost; - Reduced manual ID checks by 10%, and created dashboards cutting detection time from 2 days to hours, reducing backlog by 85%.

- Led analytics stream in the reengineering of the credit decision system (Blaze Advisor, FICO) covering POS, Cash, and Card products; - Acted as process consultant in an Agile project launching online loans; - Contributed to a unified “multi-offer” approval flow improving product cross-selling and risk segmentation; - Delivered SQL-based analytics and ad-hoc reports enhancing decision accuracy and operational control.

- Maintained and configured the credit decisioning system (Blaze Advisor, FICO) in compliance with internal risk policies; - Developed and maintained SQL reports tracking key credit approval KPIs; - Supported release testing, issue localization, and parameter tuning in production environments; - Assisted in business rule updates for retail loan products, ensuring stability and consistency of decision logic.