Search technologies are the DNA of Yandex's search business group. Currently, one in five queries is related to product search, a scenario that accounts for 40% of profit. We are developing a search engine that scours tens of thousands of online stores and plan to integrate a convenient AI consultant into it. It will compare products by their characteristics and help users decide where it's best to buy a particular item.

Our team is responsible for the quality of the product database structure. We create product cards and 'stick' offers from stores to them, so that users can choose the best price or the fastest delivery.

Our database contains billions of products: ML models have to compare thousands of products per second, and we devise non-standard solutions to optimize these processes. We talked in detail about the product matching task (aka Product Matching) at the ML Party.

What tasks await you

Training models and implementing them into the final product Fine-tuning YandexGPT models for content generation, classification tasks, etc. Training neural networks and gradient boosting models. Working on clustering billions of products. Carrying out the full cycle of work from dataset collection to model deployment in the final product. Researching the latest ML approaches and proposing ideas to improve model quality. Finding the trade-off between model quality and inference speed.

We expect that you

• Have trained ML models and put them into production
• Can formulate tasks in ML terms, understand how to measure results, are familiar with various algorithms and can choose the appropriate one
• Care about the high quality of the final product
• Are ready to immerse yourself in new technologies

Will be a plus

• Have trained large ML models like BERT
• Have worked with generative networks like ChatGPT
• Have built infrastructure for ML processes
• Have proficiency in C++ or are willing to learn it

Work conditions

• Work in a strong team (our team includes graduates of the Yandex School of Data Analysis and ACM champions)
• Active career growth: we have bold projects that improve the service by multiples, not by tenths of a percent
• Challenging tasks for services with millions of users.
• Speaking at conferences, joint research projects with students of the Yandex School of Data Analysis
• The opportunity to influence the process and the result
• Market-level or above-market salary
• Bonuses every six months for everyone who successfully passes the review
• Flexible work schedule
• Mortgage programs at 3% for 10 years or interest-free for 3 years
• Meal compensation
• Extended voluntary health insurance program, 80% of the cost of insurance for spouses and children is covered
• Gym, fitness center, yoga in the office
• Free parking