We are looking for a Data Scientist

Required:

• Proficient in Python and SQL;
• Good understanding of machine learning methods and mathematical statistics: linear models, Random Forest, gradient boosting;
• Practical experience with time series: data preparation, forecasting, quality assessment, and cross-validation;
• Experience with gradient boosting libraries: XGBoost, LightGBM, or CatBoost;
• Knowledge of mathematical and discrete optimization methods: LP, MILP / MIP;
• Practical experience with at least one optimization tool: Pyomo, OR-Tools, PuLP, or similar;
• Experience using optimization solvers: Gurobi, CPLEX, SCIP, COPT, GLPK, or similar;
• Experience developing ML solutions from data analysis and model building to a working prototype;
• Proficient in Git.

Desirable:

• Experience applying advanced time series forecasting models: Prophet, LSTM, DeepAR, or similar;
• Experience with big data and data infrastructure: Spark / PySpark, S3;
• Experience integrating ML models into services and understanding of basic MLOps: FastAPI, Docker, CI/CD, Kubeflow, or similar;
• Experience writing unit and integration tests, understanding of OOP;
• Experience in FMCG / Supply Chain / Demand Planning / Forecasting / ERP / APS.

Tasks:

• Collecting, cleaning, and preparing data from various sources for further analysis;
• Conducting primary data analysis: identifying patterns and anomalies, creating visualizations;
• Developing and implementing machine learning models (including regression, classification, and clustering), with a special focus on developing time series related models;
• Developing an optimizer based on linear and mixed-integer optimization, LP, MILP;
• Evaluating model quality, tuning hyperparameters, and performing cross-validation;
• Participating in the integration of models into the company's business processes;
• Preparing analytical reports;
• Effective interaction with business representatives and technical teams;
• Improving current approaches and learning new tools and algorithms.

For quick response: