ML Developer for the Kinopoisk Recommendations Team

Kinopoisk is one of the largest online cinemas in Russia.

Our recommendation team creates ML algorithms for selecting content that users are genuinely interested in. This is how we implement Discovery scenarios and make the service more convenient: people visit Kinopoisk more often and interact with it more. To achieve this, we test and implement new ML solutions, and also help develop the product.

What tasks await you

Hypothesis formulation and A/B experimentation You will propose new hypotheses and test them — in order to delight viewers with ML models and product logic, helping them find the most interesting series and movies on Kinopoisk.

Development and advancement of ML technologies You will be responsible for collecting and preparing data, selecting ML approaches to solve product tasks, as well as formulating quality metrics — to evaluate models offline and during A/B tests.

Writing production code in Java Our recommendation serving runtime service is written in Java. At the implementation stage, you will need to write reliable and efficient production code. Our services are already quite mature — this allows us to implement complex scenarios.

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

We expect you to

• Understand the principles of classical and deep ML, and have applied them in practice
• Have prepared and processed tabular data (using SQL-like approaches)
• Know the basics of mathematical statistics
• Be capable of independently conducting research — from task formulation to presentation of results
• Be ready for the full cycle of ML solution implementation: from data processing to production deployment and maintenance of the service with the ML model
• Want to improve the product