ML Developer for Ecom Scenarios Team

Our team works on projects at the intersection of Advertising, Alice, and Ecom. Our main goal is to improve the quality of the "Find Cheaper" agent. The agent helps users find the best deals online.

If you are passionate about practical ML work, enjoy experimenting, and taking solutions to production in ambitious projects, we would be happy to meet you.

What tasks await you

Model Training You will work with large architectures (up to 8B parameters): selecting and preparing datasets, configuring training and fine-tuning, improving quality and cost metrics, experimenting with architectures and instruction sets. Your responsibilities will include research to find and develop SOTA solutions for our tasks.

Pipeline Work and Production Deployment You will build end-to-end ML pipelines: collecting and preparing data, working with offline and online metrics, and optimizing inference (quantization, distillation, caching, distillation into smaller models).

Data Annotation You will organize and oversee data annotation (e.g., assessing how relevant a product is in a text or collection): creating instructions, monitoring the quality of assessors' work, and experimenting with LLM/agent-based annotation. Quality annotation is key to the success of almost any NLP project.

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

We expect you to have

• Knowledge of Python 3 YQL
• Experience with ML/DL

It will be a plus

• Practical experience in NLP or RecSys