Yandex Go is a real-time service, and one of its key tasks is managing supply and demand (the needs of passengers and the availability of drivers). Subsidies are a tool that helps avoid strong imbalance. Our team improves Yandex Go using big data and machine learning algorithms and makes the service more accessible to millions of drivers and passengers every day.

About the Team

Learn about the development of Yandex's urban services at dev.go.yandex

What tasks await you

Managing the full ML project cycle You will be responsible for the entire project cycle: communicating with the business stakeholder and formalizing tasks, choosing the technology stack and model architecture, and implementing the ML model into production.

Collaboration with related ML teams You will jointly select and implement technologies optimal for solving tasks.

Identifying growth opportunities and improving business metrics You will formulate hypotheses and test them experimentally.

We expect you to

• Have a strong command of the classic ML development stack: Python (sklearn, XGBoost or CatBoost, Pandas, Seaborn, Matplotlib, etc.), SQL
• Understand how to properly collect data for an ML task
• Have a strong algorithmic background
• Be able to formulate hypotheses, prioritize them, and test them on data

It will be a plus if you

• Have industrial RL experience
• Have implemented and scaled neural network approaches for tabular data/sequences
• Can work with large data volumes and understand the MapReduce paradigm (Hadoop, Spark)
• Have conducted experiments and analyzed A/B test results
• Are familiar with PyTorch Lightning / Lightning frameworks and have practical experience with one of them