About
DigitStat, a limited liability company under the National Statistics Committee of the Republic of Uzbekistan, invites a specialist with strong analytical thinking skills who is ready to create modern AI and Machine Learning solutions based on statistical data to join as a Data Scientist / ML Engineer (Middle+). DigitStat was established under Resolution No. PQ-103 of the President of the Republic of Uzbekistan dated March 13, 2026, and is the single integrator for the digitalization of the national statistics system.
Responsibilities
•Develop, test, and deploy Machine Learning models based on statistical data in a production environment.
•Create time series forecasting models in macroeconomics, prices, and demographics.
•Develop classification and regression models based on data from registers and statistical observations.
•Participate in creating the STAT-AI system based on official statistical data, including working with LLM and RAG architecture.
•Develop mechanisms for indicating the source of official data used in STAT-AI responses.
•Develop AI solutions that work with statistical data in Uzbek and Russian.
•Develop rapid prototypes of AI/ML solutions for the needs of government agencies and corporate clients.
•Carry out feature engineering, model evaluation, and data preparation.
•Work with imbalanced, incomplete, and variable-quality real-world data.
•Manage several Data Science and AI projects simultaneously.
Requirements
•At least 3 years of practical work experience in Data Science or Machine Learning.
•Confident working knowledge of Python and SQL (PostgreSQL).
•Good knowledge of the Pandas, NumPy, and Scikit-learn libraries.
•Practical experience working with PyTorch or TensorFlow.
•Good understanding of basic statistical methods, including sampling, confidence intervals, hypothesis testing, and time series.
•Practical experience in feature engineering and evaluating ML models (model evaluation).
•Ability to process imbalanced and "dirty" data.
•Experience with Git and understanding of the process of deploying ML models to a production environment.
•Readiness to work with confidential data, such as statistical secrets and personal data, under strict regulations.
•Fluency in Uzbek and Russian.
Nice to have
•Experience with LLM, Hugging Face, RAG, LangChain/LangGraph, AI Agents, and Tool Calling technologies.
•Knowledge and experience with LoRA/PEFT and optimizing models for CUDA/GPU.
•Experience in Uzbek-language NLP projects, including tokenization and fine-tuning models.
•Experience with GeoPandas, PostGIS, and QGIS.
•Experience with ClickHouse or other analytical databases.
•Knowledge of FastAPI, Docker, MLflow, Airflow, or dbt technologies.
•Understanding of official statistics methodology, including IFUT (OKED), registers, and census processes.
•Experience working on Data Science/AI projects in the public sector or at research companies.
•Understanding of requirements for AI systems in the public sector, including human-in-the-loop and the explainability of model decisions.
Conditions
•Formal employment in accordance with labor legislation.
•Competitive salary in line with the IT market — the final amount will be determined based on interview results.
•Additional bonuses based on the results of completed projects.
•Direct participation in creating AI and Machine Learning products for the national statistics system.
•Work on projects in modern areas such as STAT-AI, forecasting, NLP, and geoanalytics.
•Develop professional experience using large volumes of real statistical data.
•Work in collaboration with a professional Data Science, Data Engineering, and AI team.
•Work schedule: 5/2, 9:00–18:00.
•Work location: 63 Mustaqillik Avenue, Tashkent — National Statistics Committee building.
•A 3-month probation period, with an initial 14-day onboarding program.
