NLP Developer for Browser

Our team is responsible for integrating and improving Alice in Yandex Browser, developing and enhancing ML in browser functions, enabling the AI assistant to see web pages and interact with them. We have launched short summaries of videos and articles, capabilities for neuro-editing (working with text based on YandexGPT), and a feature to answer questions about a page. We are looking for strong ML engineers who will develop and create solutions based on both modern agent architectures and LLMs, as well as classical machine learning models.

What tasks await you

Improving current models and training new ones You will be involved in training both classifiers and standard generative models, as well as working on improving complex scenarios based on agents and complex pipelines at the intersection of several technologies.

Our tasks include collecting and analyzing data for fine-tuning, as well as the training process itself using SFT and RL approaches.

Building quick prototypes for new projects To launch new products, you need to be able to quickly test hypotheses and build MVPs. The current main focus is on scenarios involving agents and using RAG for rapid processing of large data contexts. It is important for us to quickly test various hypotheses and be able to assemble prototypes quickly, through which we determine future projects and our capabilities.

Automation of ML model validation You will implement systems for the automatic evaluation of model performance in production.

We expect that you

• Have experience in ML
• Have good knowledge of Python
• Are proficient in classical algorithms and models
• Have a good understanding of NLP and LLMs

It will be a plus

• Have experience working with agents and RAG pipelines