ML Developer for Safe Search Team

ML Developer for Safe Search Team

Our team ensures that unwanted content does not appear in search results: making search completely safe and interesting for people of any age. We use a wide stack of ML tools: from boosting to neural networks. We do not engage in abstract research, but actually change the product and manage projects from start to finish. Join us!

What tasks await you

Searching and Formalizing Problems You will work with big data using MapReduce: find search problems and formalize them as offline and online metrics, explain why links or documents appeared in search results or disappeared from them, and propose a way to fix the situation.

Developing Instructions You will need to create and check the relevance of instructions for crowdsourcing.

Training and Implementing Models You will train neural and CatBoost models, implement them into search runtime. You will experiment with training LLM and VLM, apply models to change search ranking, and conduct A/B experiments.

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

We expect that you

• Have experience in data analysis
• Have developed in Python or C++
• Are familiar with classical ML
• Are ready to delve into complex search infrastructure
• Are focused on implementation and ready to manually check the result of your work

Working Conditions

• Modern office in the center of Moscow with parking and everything needed for comfortable work and rest
• Official employment
• A badge that can be used to pay for meals on the campus and in its vicinity
• Flexible schedule: we do not control who comes and goes when, the main thing is to complete tasks
• Extended voluntary health insurance program with dentistry, examinations, home doctor calls, payment of 80% of the VHI cost for spouses and children
• High income and bonuses every six months for everyone who works effectively
• Opportunity to participate in Yandex educational programs, lectures, and meetups
• Work in a talented expert team where you can learn a lot
• Culture of openness and mutual assistance
• Opportunity to quickly see work results and create services used by millions of people