We are looking for a strong researcher for Yandex's main product — Search.

Our team handles the semantic part of Search. We:

• Help Search better understand user queries, web documents, and their relationships
• Work at the intersection of NLP and ranking
• Help build Yandex's agent systems
• Use hundreds of GPUs for experiments
• Influence tens of millions of people with our releases

What tasks await you

Find growth points for Search quality factors You will need to identify growth areas for signals and models and formulate research hypotheses based on them.

Close the found growth points through data You will order labeling, create and validate datasets, and improve the quality of training samples.

Work on the quality of LLM-based models You will enhance the quality of LLM models by enriching context and improving architecture.

Develop ML components end-to-end You will be responsible for the full development cycle — from idea and experiment to integration into a mature production service.

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

We expect that you

• Have deep expertise in at least one of these areas: NLP, search, or recommendations
• Know how to do practical ML system design
• Confidently write in Python, work with SQL and PyTorch
• Have a good foundation in probability theory and mathematical statistics
• Are ready to challenge the status quo and see problems in established solutions
• Quickly learn new technologies
• Clearly convey your thoughts to the team and stakeholders
• Demonstrate strong ownership over your area of responsibility