Skip to content
AG

Alibek Galiyev

@Aliba24

ML Engineer

Data Scientist and Risk Manager based in Almaty, specializing in credit-risk modeling, backend services (FastAPI), and ML pipelines. Passionate about applying PyTorch to complex datasets. When not in front of laptop, I like hybrid training and basketball.

Career

Experience: 13 years

Experience

Head of Risk / Lead Data Scientist

FixPay

Fintech startup doing POS lending for Avito, Wildberries, and M-Video.

Sep 2025 — Present

Leads risk modeling and decisioning for POS lending, from model development and provider testing to portfolio and anti-fraud analysis.

• Build predictive models from scratch using advanced ML architectures (gradient boosting and neural networks), managing data preparation, feature engineering, training, deployment, and monitoring (PSI/CSI, performance decay). • Lead R&D with external data providers, prototyping models to measure incremental Gini lift before integration. • Manage the full decisioning pipeline in Axiomatika: application flow, validation and bureau checks, and effective-rate assignment logic.

Senior Data Scientist

Freedom Finance Global PLC

Dec 2024 — Sep 2025

Developed retail customer segmentation, advisory, and recommendation models, and established retraining pipelines to address model drift.

• Developed ML models for customer segmentation and behavior analysis across the brokerage's retail customer base. • Built advisory and recommendation systems and deployed predictive models to production. • Partnered with stakeholders to translate commercial questions into scoped ML use cases and established retraining pipelines to combat model drift.

Head of Risk / ML Engineer

LFTech

Apr 2022 — May 2024

Led risk modeling and data infrastructure, with OpenScoring models handling more than 50,000 loan application decisions daily.

• Designed and maintained production ML models, scoring 50,000+ daily loan applications via OpenScoring. • Developed anti-fraud ML models utilizing JuicyScore device fingerprinting features. • Established the risk data warehouse and owned ETL/ELT pipelines feeding analytics and reporting.

Head of Risk

SmartFinance

Aug 2020 — Apr 2022

Led risk work on alternative-data scoring models and optimized credit and anti-fraud rules in the loan origination flow.

• Developed and monitored scoring models leveraging alternative data sources. • Optimized credit and anti-fraud rules across the loan origination flow, utilizing data-driven approaches.

Head of R&D

Silkway Ventures

Feb 2018 — Apr 2020

Led loan portfolio model development and built the data warehouse and management reporting systems.

• Led ML model development (Logistic Regression, Random Forest, XGBoost) for predictive analytics on loan portfolios. • Architected the data warehouse and management information systems supporting daily reporting.

Education

Master's, Reservoir Engineering

INPL

2013

Master's, Mechanics

Al-Farabi Kazakh National University (KazNU)

2013

BTech, Mathematical & Computer Modeling

Al-Farabi Kazakh National University (KazNU)

2011

Selected projects

Daily ETL Pipeline and Automated Reporting

Engineered a daily Talend ETL pipeline extracting and transforming data from approximately 30 production credit conveyor tables, loading the central DWH and generating daily business reports in Power BI.

scorengine

Developed an open-source Python library that automates recurring ML and scorecard tasks, including exploratory data analysis, IV/WoE binning, feature selection, and validation, to standardize scoring workflows.

FinKarta Enrichment Microservice

Built a standalone FastAPI and Pydantic service that queries FinKarta enrichment APIs and runs in Docker, enabling downstream systems to enrich profiles with a single call.

Production Scoring Pipeline on OpenScoring

Built and deployed PMML scorecards to a Java-based OpenScoring service at LFTech, supporting more than 50,000 daily loan application decisions.

Skills

PythonSQLGoSwiftGroovyRDeep LearningPyTorchGradient BoostingXGBoostLightGBMCatBoostSymbolic MathSymPyLogistic RegressionRandom ForestClusteringUplift Modelingscikit-learnpandas