Our current challenge is to create the most advanced digital bank in Uzbekistan. We are looking for a Data Scientist.
Responsibilities:
- Features and Feature Platform: design, develop, and validate features; ensure their correct availability in training and inference (offline/online), maintain documentation and feature lifecycle.
- Feature Quality: evaluate the significance, stability, and impact of features (statistics, leakage, drift), optimize existing ones and add new ones.
- End-to-end Models: develop and improve ML models, conduct experiments, validation, and interpretation of results.
- Shipment to Production: bring the solution to production level together with ML/Engineering: pipelines, CI/CD, MLOps, monitoring of quality and data after release.
Requirements:
Feature Platform / Feature Engineering
- Understands what a feature is in product and technical logic: definition, aggregation window, granularity, entity key, update rules, offline/online.
- Knows how to develop features in a platform/feature store (Feast or similar): data sources, materialization, quality control, versioning, and repeatability of calculations.
- Aware of typical risks: data leakage, bias, duplicates, key inconsistencies, time shifts, correct handling of time-aware data.
Modeling
- Confidently builds ML models on tabular data: problem definition, target labeling, validation (CV/Time-split), metric selection, tuning.
- Can explain factor contributions and make decisions on features/models based on data (feature importance, SHAP/similar, ablation, stability by segments).
Technical Stack: Python + SQL + DWH
- Strong Python for production code: pandas/NumPy, scikit-learn, CatBoost/LightGBM (or similar), clean project structure, testability.
- Strong SQL: complex JOINs, window functions, aggregations, query optimization; confident work with DWH (Greenplum/PostgreSQL).
- Understands how to build reproducible datasets and calculations on big data (performance, profiling, resource control).
Production Mindset: CI/CD and MLOps
- Understands the full ML cycle: data preparation → training → artifacts → deployment → inference → monitoring.
- Has experience/understanding: Git, CI/CD, Docker, MLflow (or similar), basic monitoring (model quality, drift, data quality, latency).
Bonus Points:
- Practice in marketing and customer behavior tasks: churn/retention, uplift, targeting, segmentation, LTV, A/B thinking.
- Experience in product analytics and communication: ability to formulate hypotheses and translate analysis results into concrete business solutions.
Conditions:
- Employment in accordance with the Labor Code of the Republic of Uzbekistan.
- Paid vacation and sick leave.
- Relocation assistance for those moving to sunny Tashkent.
- High competitive salary.
- Ambitious project where you will play one of the most important roles.
Work Location:
Tashkent, Nukus Street, 29A
Interested in the vacancy? Send your resume to: recruiter contact
