Team Lead
Head of the "Problem Goods" direction in anti-fraud. Scaling the team and product line, responsible for business metrics and technological development.
- Built and growing a team of ML engineers — hiring, mentoring, code review, engineering culture
- Scaled the product line — launched new services for detecting fraudulent schemes
- Implemented MLOps practices (ClearML, MLflow, Airflow), automated deployment and model monitoring
- Coordinate work with product teams, analytics, and related directions
- Participate in strategic planning and anti-fraud roadmap formation
Python · OpenAI API · LLM · FastAPI · Kafka · ClickHouse · S3 · Airflow · ClearML · MLflow · GitLab CI/CD · Grafana · Prometheus
Senior ML Engineer
One of the first members of the new "Problem Goods" direction in anti-fraud. Built from scratch ML/LLM solutions for detecting and blocking problematic goods on the marketplace.
- Launched the direction from scratch — from requirements gathering to production deployment
- Developed 10+ production services for combating fraud
- Applied a hybrid approach: LLM, classical ML models, statistical methods, and heuristics
- Built integration via Kafka and FastAPI between services
- Developed Grafana dashboards for tracking model quality and business metrics
Python · LLM · FastAPI · Kafka · ClickHouse · S3 · Airflow · ClearML · GitLab CI/CD · Grafana · Scikit-learn
Instructor
Instructor on the Machine Learning Engineer course.
ML Engineer
SafeData — a high-load service for profiling and anonymizing personal data in bank databases.
- Trained transformers (BERT) for sequence classification and NER tasks
- Developed a data generator based on fixed structure and LLM
- Wrote and optimized algorithms and regular expressions for finding PII in DB columns
- Developed atomic services for deployment in bank infrastructure
- Product is used by the entire bank and subsidiaries; macro F1 across 100+ domains > 90%
Python · Regex · PyTorch · Transformers · FastAPI · Docker · Jenkins · Airflow · MLflow · OpenShift
ML Engineer
Developed and implemented ML solutions for oil production optimization and automation of engineering processes.
- Developed and deployed to production ML tools based on LSTM networks for predicting well parameters
- Designed data pipelines from MySQL databases of field well data
- Automated business processes for engineers using Python and VBA tools
- Published papers in Russian and English, presented at 20+ specialized conferences
Python · LSTM · Scikit-learn · MySQL · SQL · VBA · Pandas · NumPy
