Data Scientist for Yandex Cloud ML Services

Our team develops Yandex Cloud ML services: Yandex SpeechKit, OCR and Vision, Translate, YandexGPT, Alice AI, and tools that help Data Science specialists train neural network models on tens of terabytes of data and hundreds of GPUs, and control the deployment of these models both in the cloud and on the client's premises.

In addition to these core ML technologies, we also develop products on top of them. For example, Realtime API (for speech-to-speech voice assistants), AI Search (a platform for building RAG solutions), SpeechSense (a speech analytics platform).

We are looking for a Data Science and analytics specialist.

Read more about our team on the Yandex Cloud ML services page.

About the team

Subscribe to the Inside Yandex Cloud Telegram channel to learn more about our team and technologies!

What tasks await you

Analyze user experience and build metrics and processes for evaluating the quality of ML services and products To improve the quality of our models and products, understand their strengths and weaknesses, and prioritize plans for developing new functionality, high-quality qualitative and quantitative analytics are essential. We cannot live without metrics and instruments that show which direction to move in—and you will help us build such instruments and the processes for their regular updating.

Build data labeling processes for training ML models One of the key factors affecting the quality of ML models is the training data. You will develop processes for collecting and labeling data, evaluate and improve the quality of these processes. Examples of tasks: labeling data for speech recognition in Uzbek; collecting training data for text classification tasks.

Help validate product hypotheses We not only develop models but also help implement them into products. During implementation, many hypotheses arise about what can be done with the models and how. To validate hypotheses, it's often necessary to develop and analyze a prototype, choose the best solution option. You will help product managers and developers build prototypes and explore technical solution options. Example task: analyzing different text clustering algorithms based on GPT.

More about analytics at Yandex — in the channel Yandex for Analytics

We expect you

• Write in Python
• Have a good knowledge of statistics
• Understand modern ML and NLP methods
• Have processed and analyzed large volumes of data using pandas, SQL, statistical packages, data visualization libraries, Spark SQL, Spark, Hadoop
• Have worked with version control systems, e.g., Git

Will be a plus

• Have worked with DataLens
• Have worked with crowdsourcing (Yandex Toloka, Yandex Tasks)