Аркадия 2 часа назад
Senior Applied ML Engineer (Python)
- Удалённо
- Опыт от 6 лет
- Полная занятость
PythonSQLPostgreSQLDockerGitC#Machine LearningMLA/B testingPyTorchpandasNumPyTensorFlowScikit-learn
Описание
Responsibilities
- Development of ML models: predicting the probability of purchasing additional services at different prices, optimizing pricing decisions, building neural network models
- Experimental methodology: designing controlled exploration within an acceptable price range, randomization rules, assignment probabilities for pricing options, and safety constraints
- A/B testing: designing controlled experiments to compare new pricing policies with current ones, defining primary and guardrail metrics, stopping criteria, and rollout conditions
- Full machine learning lifecycle: training, validation, inference, periodic retraining, model versioning, feedback loop between predictions, decisions, and actual outcomes
- Development: implementation of an end-to-end ML prototype, including minimal HTTP API, data storage, and simulation of an external booking system
- Logging and data: correct logging of assigned prices, experimental variants, assignment probabilities, purchase facts, and business metrics
Requirements
- Proficient in Python, NumPy, pandas, scikit-learn, PyTorch or TensorFlow, as well as C# for backend integration
- Experience in developing, training, and applying ML models, including neural networks, feature engineering, and working with tabular data
- Understanding of the mathematical foundations of machine learning, probability theory, statistics, optimization methods, and A/B testing principles
- Ability to design controlled exploration, randomization rules, primary and guardrail metrics for pricing experiments
- Experience in organizing the full machine learning lifecycle: training, validation, inference, retraining, versioning, and feedback
- Proficient in SQL, PostgreSQL, Git, Docker, and skills in developing Python applications with autotests
Nice to have
- Experience in dynamic pricing and revenue optimization: demand forecasting, price elasticity, willingness to pay
- Familiarity with advanced methods of causal inference and sequential decision making: contextual bandits, off-policy evaluation, IPW, doubly robust estimation
- Familiarity with MLOps tools (MLflow, orchestration, model monitoring) and experience in operating ML services
- Experience in the aviation industry, e-commerce, or revenue management
We offer
- Full-time employment
- Remote work
- 5/2 schedule