ML Developer for the AI Search Quality Team

The world is rapidly changing: more and more search scenarios are being covered by conversational AI, and we are actively participating in this race. Modern AI, for example, Alice, not only answers from memory but also actively searches for external sources from the internet—and the quality of the provided sources is critical for a good response.

Our E2E team is responsible for the quality of search for AI. Join us and help invent the search of the future!

What tasks await you

Improving Retrieval Quality Currently, our AI services work on top of the traditional Yandex Search with minor improvements. Search is heavily optimized for human consumption with all its peculiarities. You will significantly rework the search stack for AI consumption to provide the most relevant and expert sources for every query. Together with our analytics team, you will analyze losses and weak spots, develop new quality signals, and implement models.

Improving Context Engineering Finding documents is not enough—you need to prepare good context from them for the LLM. There are many open challenges here: how to extract only the important and necessary information without breaking the meaning and while fitting strict production constraints? You will have to find the answers.

Accelerating Infrastructure and Saving Costs Search is expensive. We spend a lot on hardware, and the RPS from agent systems is higher than from humans. Together with the infrastructure team, you will optimize the search stack without losing quality.

Learn more about ML at Yandex — in the Yandex for ML channel

We expect that you

• Have strong engineering and ML culture, have applied ML in production
• Possess a broad outlook and knowledge of modern NLP
• Understand that data is the most important thing in ML, are ready to look at it extensively and attentively
• Can critically evaluate metrics and model results
• Want to see the product and the user behind the numbers
• Are independent and responsible