Чтобы адаптировать резюме под вакансию или составить сопроводительное письмо, загрузи резюме
О рекламодателе
ОБЩЕСТВО С ОГРАНИЧЕННОЙ ОТВЕТСТВЕННОСТЬЮ "ЦЕНТР НАЦИОНАЛЬНЫХ ИНТЕЛЛЕКТУАЛЬНЫХ СИСТЕМ" ИНН: 9704271170
описание
The team designs, builds, and optimizes recommendation and ranking systems at scale, developing models that power personalized experiences and delivering machine learning solutions in a modern cloud environment.
задачи
Design and build recommendation and ranking models, including learning-to-rank, re-ranking, gradient-boosted, and neural rankers
Optimize models for multiple objectives and analyze Pareto fronts to balance competing goals
Evaluate recommender systems offline using ranking metrics such as NDCG and recall@k
Address position and selection bias in logged data to ensure robust model performance
Apply statistical rigor through significance testing, confidence intervals, and A/B test and holdout design
Process and analyze large-scale behavioral and clickstream data
Develop, test, and maintain clean, version-controlled code using Git
Track experiments and manage models throughout their lifecycle
Deploy and manage machine learning workflows on a cloud data platform
Collaborate with cross-functional teams to deliver production-ready ML solutions
требования
3+ Years of hands-on experience building recommendation or ranking models
Understanding of multi-objective optimization, including Pareto-front analysis
Expertise in offline recommender evaluation using ranking metrics such as NDCG and recall@k
Familiarity with position or selection bias in logged data
Strong statistics background, including significance testing, confidence intervals, and A/B test and holdout design
Proficiency in Python and SQL
Skills in large-scale data processing with Spark, pandas, and polars
Competency in ML libraries such as scikit-learn, PyTorch, and LightGBM/XGBoost
Experience with behavioral or clickstream data
Knowledge of experiment tracking and model management tools such as MLflow
Familiarity with cloud data platforms, preferably Databricks
Ability to write clean, tested, version-controlled code using Git
English proficiency at Upper-Intermediate level (B2) or higher
Будет плюсом: E-commerce or retail background, understanding of counterfactual or off-policy evaluation, familiarity with contextual bandits or online learning, experience with real-time model serving
.svg)
