IT_ONE 4 дня назад

ML Engineer

  • Удалённо
  • Полная занятость
PythonSQLpandasNumPyScikit-learnLGBMCatBoostXGBooststatsmodelsProphetSkforecastSktime

Описание

Responsibilities

  • Develop production models for business tasks — forecasting, classification, anomaly detection, time series analysis
  • Participate in the full cycle of ML projects — from setting the task with the customer and preparing data and features to training, validating models, deploying to production, and assessing impact on business metrics
  • Design and train models on a classic ML stack (gradient boosting, linear models, ensemble methods)
  • Develop data processing pipelines, feature engineering, and model validation
  • Assess the risks of using ML solutions, considering the specifics of business and technological processes
  • Research modern approaches in tabular ML, forecasting, and anomaly detection

Requirements

  • Confident knowledge of Python
  • Proficiency in Pandas, NumPy, Scikit-learn
  • Experience with XGBoost/LightGBM/CatBoost
  • Experience with time series and forecasting — Statsmodels, Prophet/Sktime/Skforecast
  • Knowledge of classic ML methods — decision trees, boosting, linear models, clustering, dimensionality reduction
  • Statistics and interpretation of experiments (A/B, hold-out, cross-validation)
  • SQL for independent data work
  • Git
  • Experience deploying ML models to production and building reproducible experiments
  • Higher technical education (Computer Science, Applied Mathematics, AI/ML, or related fields)
  • Experience using Apache Spark (PySpark) for preparing and processing large volumes of data
  • Product thinking — understanding why a model is needed by the user and what metric it affects
  • Willingness to seek growth opportunities, propose non-standard solutions, and defend them in meetings with the customer
  • Good communication and teamwork skills

Nice to have

  • DVC, MLflow, Airflow, Docker/Kubernetes

We offer

  • Remote work
  • Full-time employment

Hiring process

  • Apply at •••