ML Developer for Voice Input Applications Team

We are the voice input applications team, working on speech recognition development for various applications within Yandex. We taught Alice to hear speech in other languages, built voice input for Yandex Translate, Yandex Keyboard, Yandex Talk, developed ASR for video translation in the browser, and are constantly improving these technologies. Our team also often takes on ambitious R&D projects that go beyond voice input.

We value experience in ML and an interest in research, and we are ready to teach you how to work with audio.

What tasks await you

Developing ASR for Yandex Talk Yandex Talk is an application that simplifies communication between people with hearing impairments and hearing people. We want to add new capabilities to our model: implement speaker change detection, support multilingualism, improve speech recognition for speech impairments, in difficult acoustic conditions, and for special terminology.

Improving ASR quality for Yandex Keyboard Yandex Keyboard allows millions of users to input text faster, including through voice input, and our task is to make voice input even more convenient and better. You will experiment with architectures, training approaches, and also teach the model to predict punctuation.

We expect you to have

• Worked in the field of DL, for example with CV or NLP, know PyTorch
• Understand the principles of machine learning
• Be well acquainted with Python

It will be a plus if you have

• Worked with ML in the field of voice technologies
• Worked with C++
• Worked with large volumes of data