Moscow, Russia
Email: lebedev.aale@gmail.com
LinkedIn: https://www.linkedin.com/in/lebedev-andrey
GitHub: https://github.com/lebedevaale
Retail portfolio analyst at Sber, author-lecturer at Central University, and researcher at Higher School of Economics.
Expert in ML for credit risk forecasting (PD/LGD), data engineering (Polars, PySpark, Greenplum), and time series anomaly detection.
PhD candidate researching ML for complex systems on chaos edge (thesis 2027).
- Languages & ML: Python 3 (Polars, PySpark, sklearn, LightGBM, XGBoost, CatBoost, PyTorch, statsmodels), SQL, Matlab
- Data Platforms: Greenplum, Hadoop, PostgreSQL, Oracle DB, MS SQL, ClickHouse, Neo4j
- Tools: Dash, Plotly, Flask, Selenium, SciPy, nolds, filterpy
- Domains: Credit risk modeling, fraud detection, time series anomalies, ETL automation, teaching SQL/Data Engineering
Sber · Moscow, Russia (Office)
Aug 2025 – Present
- Developed ML models for long-term PD and reserves forecasting for consumer loans, mortgages, credit cards, and car loans using Python (Polars) + SQL (Greenplum) — reduced forecast error by 60% on average.
- Managed development of ML models for PD and LGD forecasting across retail loan products.
- Automated and refactored analytical pipelines in Python (Polars/PySpark) + SQL (Greenplum/Hadoop) — relieved product analysts of 50% ad-hoc workload.
Stack: Python 3 (Polars, PySpark, sklearn, LightGBM, XGBoost, CatBoost, statsmodels, Dash, Plotly, Flask), Greenplum, Hadoop
Central University · Moscow, Russia (Hybrid)
Aug 2025 – Present
- Co-authored SQL and Data Engineering courses — developed 1/3 of all lecture and seminar materials.
- Delivered lectures and seminars on data engineering topics.
Stack: Python 3 (PySpark), PostgreSQL, Greenplum, Hadoop
Higher School of Economics · Moscow, Russia (Remote)
Apr 2022 – Present
- Conducted ML research on anomalies in financial, physical, and social time series using statistical and physical tools in Python and Matlab.
- PhD thesis "Machine Learning in Analyzing Performance of Precursors to Self-Organisation in Complex Systems on the Edge of Chaos" to be presented in 2026.
- Published 5 WoS/Scopus-indexed articles (4 in Q1/Q2 journals).
Stack: Python 3 (Polars, sklearn, LightGBM, XGBoost, CatBoost, PyTorch, statsmodels, Dash, Plotly, Flask, Selenium, SciPy, nolds, filterpy), PostgreSQL, Neo4j, ClickHouse
PJSC MTS-Bank · Moscow, Russia (Hybrid)
Feb 2025 – Jul 2025 (6 months)
- Developed ML model based on mid-term session/transaction activity to analyze call center incidents — accelerated separation of real fraud from false positives, reducing response time.
- Automated Polars-based DA pipelines, saving 10+ hours/week on fraud analysis; fully automated fraud/rule failure parsing with daily recommendations.
- Mentored 3 junior analysts in Python, SQL, and Git.
Stack: Python 3 (Polars, LightGBM, XGBoost, CatBoost, statsmodels, sklearn, Dash, Plotly, Flask, Selenium), Oracle DB, PostgreSQL
Bank SOYUZ · Moscow, Russia (Hybrid)
Mar 2023 – Feb 2025 (2 years)
- Developed/validated annual and quarterly ML models for corporate borrower default prediction using financial statements and company characteristics.
- Automated data collection from APIs/open sources via parsers, redirecting corporate risk team to new analytical tasks.
- Technical leadership of 3 junior/middle analysts; established 2-week sprint task distribution.
- Automated Board/Central Bank capital adequacy risk reporting using Dash (ETL → PDF generation → BI server upload).
Stack: Python 3 (Polars, statsmodels, sklearn, Dash, Plotly, Flask, Selenium), Oracle DB, MS SQL
PhD Candidate (Machine Learning, Complex Systems)
Higher School of Economics · Moscow, Russia
2024 – 2027 (expected)
Master (Strategic Corporate Finance)
Higher School of Economics · Moscow, Russia
2022 – 2024
Bachelor (Business Informatics)
Higher School of Economics · Moscow, Russia
2018 – 2022
Frontiers in Physics (Q2, 2025) - A.Dmitriev, A.Lebedev, V.Kornilov, V.Dmitriev - «Self-organization of the stock exchange to the edge of a phase transition: empirical and theoretical studies»
Complexity (Q1, 2023) - A.Dmitriev, A.Lebedev, V.Kornilov, V.Dmitriev – «Twitter Self-Organization to the Edge of a Phase Transition: Discrete-Time Model and Effective Early Warning Signals in Phase Space»
Frontiers in Physics (Q2, 2023) - A.Dmitriev, A.Lebedev, V.Kornilov, V.Dmitriev – «Effective precursors for self-organization of complex systems into a critical state based on dynamic series data»
Complexity (Q1, 2022) - A.Dmitriev, A.Lebedev, V.Kornilov, V.Dmitriev – «Multifractal Early Warning Signals about Sudden Changes in the Stock Exchange States»
2026 - A.Dmitirev, I.Shulman, A.Lebedev - «Sandpile Cellular Automata with Manna Rule on the Chung-Lu Graph: Application to X Social Network»
2023 (Q4) - A.Dmitirev, A.Lebedev, N.Abbas, V.Dmitriev, V.Kornilov – «Continuum model of an avalanche-like spread of information on Twitter»
- Russian – Native
- English – C1
Last updated: March 2026