Analyst for the Search and Recommendations Team at Yandex Food

We are the search and recommendations team. We are responsible for the 'brain' of Yandex Food: we decide the order in which restaurants are shown, what to offer a user for breakfast, and how to help find needed products in the vast retail catalog. We work at the intersection of product and ML, turning terabytes of data into personalized results for millions of users.

We are looking for a strong analyst who will become a full-fledged partner for ML developers and product managers. You will not just evaluate results, but create the foundation for model training, develop data infrastructure, and critically assess (challenge) any product ideas. We will only launch what is truly useful to the user.

About the Team

How Artificial Intelligence Helps Guests Find a Restaurant - Vasily Rubtsov

What Tasks Await You

Data Architecture for ML and Recommendations Recommendations depend on the quality of the data used to train models. You will develop infrastructure for delivering data to models, search for new features and signals to improve algorithm performance. You will also help the ML team create training datasets and check model quality before launching them to production.

Validation of Product Hypotheses You will critically assess product ideas, testing them with data and common sense, not taking tasks 'on faith'. You will need to generate initiatives to improve ranking, propose new interface mechanics, and drive product development.

Development of Metrics and Quality Assessment System How do you understand if the recommendation feed has improved? You will develop sensitive online and offline metrics that show how algorithms impact user satisfaction and service economics. You will create complex experiments and figure out why some models succeed and others do not.

Ranking Management Not everything is solved with neural networks. In some scenarios, you will act as the product owner: develop and configure analytical rules and heuristics where using ML is premature or excessive. Your task is to find a balance between the 'magic' of algorithms and transparent business logic.

We Expect You To Have

• At least two years of experience in analytics (preferably in e-commerce, streaming, or marketplaces)
• Proficient in SQL (Hadoop, ClickHouse) and working with complex data structures
• Understand the principles of ML and recommender systems, know what a model needs for training and how to evaluate its quality
• Are not afraid to argue with the business and can substantiate your point of view with data
• Have excellent knowledge of mathematical statistics and conduct A/B tests (from design to interpretation of results)

Will be a plus

• Write in Python (Pandas, NumPy) for data analysis and automation of routine tasks
• Have built data marts or worked with data engineering
• Can 'sell' complex analytical solutions to stakeholders and create clear visualizations