ML Engineer / Data Scientist to the ML process automation team

Key Skills

Computer Vision (CV), Data Analysis, Data science techniques, Docker, IT, Jupyter Notebook, Machine learning techniques, Natural Language Processing (NLP), NumPy, Pandas, PyTorch, Scikit-learn, TensorFlow, XGBoost

Job Description

We develop and implement both special technical solutions for automating ML processes, as well as models and tools that allow us to make VKontakte a convenient and safe platform for everyone. We strive to ensure that our services exceed user expectations and continuously improve UX.

We are looking for a specialist who will strengthen our team and help in the implementation of large-scale projects.

You will:

• develop models and pipelines for detecting anomalies in user behavior and technical infrastructure components;
• improve the performance of existing models and fully control their revision;
• implement advanced technologies in the field of automation of machine learning and artificial intelligence.

We are interested because:

• many tasks await you from various areas of machine learning and data analysis;
• there are petabytes of data from the CV, NLP, Voice Technologies and behavioral analytics domains;
• you will have the freedom to choose technologies and ways to implement ideas;
• we have fast development cycles, frequent releases and a minimum of bureaucracy.

We expect that you:

• have received a higher technical education (with a mandatory IT profile) or are still a senior student;
• understand probability theory, mathematical statistics, linear algebra, dimensionality reduction methods and information retrieval in heterogeneous data;
• worked with frameworks and libraries - scikit-learn, Pandas/NumPy, XGBoost/CatBoost/LightGBM, TensorFlow/PyTorch;
• understand the principles of operation of automated continuous training pipelines;
• know how to set up environments and work with Docker and Jupyter Notebook.

We invite a specialist who can visit an office in Moscow or St. Petersburg, work in a combined mode or remotely. We are waiting for your responses. Good luck!