Expertise Analytics Lead in Search

Expertise Analytics Lead in Search

The core quality metric for Yandex's Proxima search results page is comprised of scores from individual signals: relevance, source authority, and others. Proxima sets the goal for search algorithms — to rank documents by query in Yandex Search. We are constantly improving the key metrics of Search, and the Offline Metrics team currently has many tasks.

One of them is to understand what constitutes a good source of an answer for the user. Search runtime processes queries and finds the most suitable sources among hundreds of billions of pages. But how do we determine how high-quality and expert these pages are?

About the team

We are a team that works on tasks without ready-made solutions. We openly discuss ideas and question any statement — the strength of the argument is what matters, not the job title. We value intelligence, commitment, and curiosity.

What tasks await you

Evaluating the quality of source documents You will define for Yandex search services what a high-quality and low-quality source document is. You will need to deeply study the specifics of services — Search, Alice, Images, Ecom — and develop clear, measurable criteria for evaluating source quality. Your definitions will become the standard for content evaluation.

Implementing product logic You will describe product logic and implement it into production: help services formulate goals for quality growth and build processes to achieve them, create new offline metrics, labeling methodologies, and ways to validate changes, check and challenge existing hypotheses about quality, conduct research, share insights with teams to collectively raise the bar for the entire Search.

Learn more about analytics at Yandex — in the channel Yandex for Analytics

We expect you to

• Possess strong product intuition: look at a task from the perspective of the end user's benefit
• Have conducted deep data analysis using Python and SQL in product analytics to test hypotheses and make decisions
• Know probability theory and mathematical statistics
• Challenge existing approaches, argue your point of view convincingly, and are open to discussion

It will be a plus if you

• Have had mentoring experience or were in a player-coach role — can assemble a team if necessary
• Participated in olympiads in mathematics, programming, or other exact sciences
• Read professional literature in English
• Understand the basics of machine learning