ML Developer for the Client Product Team

Yandex Care is a rapidly developing insurance service that offers products to protect users. Care is already available in Yandex Afisha, Market, Delivery and allows you to quickly arrange insurance directly while ordering a service or purchasing a product.

What tasks await you

Organize the full ML project cycle You will be creating a set of risk and anti-fraud models, as well as ML pricing models for Yandex Care. You will be responsible for the entire cycle: communicating with the business customer and formalizing tasks, conducting EDA, developing a prototype, conducting A/B tests, and implementing production solutions.

Design experiments, come up with new ways to evaluate quality How to collect training data so that it is unbiased? How to evaluate the effect of using a model before implementation if its impact is spread over time? You will need to find answers to these questions.

Look for the optimal way to solve business problems and influence the service We are looking for the most effective ways to solve problems, be it ML, heuristic methods, or fixing errors in the code. It is important to understand the specifics of the product and approach the work responsibly.

More about ML at Yandex — in the channel Yandex for ML

We expect you to

• Have a higher technical education
• Have good algorithmic training
• Master the classic ML development stack: Python (scikit-learn, XGBoost or CatBoost, pandas, Seaborn, Matplotlib, etc.), SQL
• Understand how to properly collect data for an ML task
• Know how to work with large volumes of data and understand the MapReduce paradigm
• Understand how DWH is structured and data processing pipelines work in production

Will be a plus

• Be comfortable with transformers
• Have high results in ML competitions, publications, patents, or open-source projects