We are creating a community for students who want to start a career in Data Science. Club members conduct experiments, participate in conferences and publish articles together with mentors from the R&D departments of MIPT, Huawei, Samsung and other Russian and international companies.
What makes successful students different?
Completed an internship at Google Spoke at NIPS Entered MIT
Passed the session as "good" Very worried about a diploma Thinks he can't write articles
We believe that the strongest students become that way through the right environment. The mentor's support and the exchange of ideas with fellow students allows you to learn in a year what it takes others four: to formulate punchline research and scientific hypotheses, conduct experiments and prepare scientific publications.
You and your mentor talk on Zoom for 30 minutes and discuss your teamwork
The problem statement
You and a mentor discuss a scientific problem and describe the stages of research
Every week you call your mentor to discuss the results of your work and plan the next one
Publishing an article
A successful project ends with publication. The prepared text of the article is sent to a pre-selected journal or conference
What will you get in the Science Club
Publications and conferences
We have high requirements for students and mentors. This allows us to focus on publications in Q1-Q2 journals and prepare our students to participate in conferences at the NIPS level.
If you want to go to graduate or postgraduate studies in Europe and the USA, you need good publications. They are also needed to get a job in the R&D laboratories of large companies. We will help you arrange articles and resumes.
Rapid growth of skills
A mentor will help you focus on what's important and not drown in a sea of articles and approaches. You will write down clear goals and metrics from the achievement. Weekly progress control allows you to monitor development.
Communities develop science. Sharing ideas and discussing success with other members is not only fun, but also broadening. Now I can't believe it, but while communicating with the students of the club, you talk to future major scientists and laboratory leaders.
You will learn how to: - Formulate research hypotheses; - Simulate problem solution and implement SotA results; - Describe research results for publication; - Review the scientific field; - Set up computational experiments
What we expect from prospective students:
Desires to explore new things
We believe that a motivated student can figure out anything, even the most difficult task.
Opportunities to find time
We expect you to be able to devote about 20 hours a week to work. It takes a long time to do a good research.
You write research code quickly, know how to use frameworks, document code, and can form research hypotheses after diving into a new field.
You have a habit of googling before asking a question. You offer solutions, you do not wait for the ready-made. It is interesting for you to understand the problem yourself, formulated in general form without detailed TR.
You ask questions. You want to discuss your own ideas with your mentor. You are interested in the area, so you yourself find new information on the topic and bring it up for discussion.
Experience of internships, hackathons, competitions at Kaggle - plus when selecting a group. Tell us about your successes.
Valentin worked in the DeepPavlov team, now he is engaged in research in the NLP division of Huawei.
Head of the Lab. neural systems and deep learning and the DeepPavlov project
Supervised the NTI project "NeuroIntellect iPavlov" and the DREAM team. Chief organizer of NeurIPS. Mikhail's publications in Nature, Artificial Life, Lecture Notes in Computer Science series, etc.
Researcher Lab. neural systems and deep learning
Winner of the competition for the development of dialogue systems NIPS Conversational Intelligence Challenge 2017. Interests: language models, coreference resolution, BERT, question-answer models for SQuAD.
Head of MIL Team
Alexey's PhD thesis is built around the topic of dynamic alignment of space-time objects.
Developer of the Samsung Artificial Intelligence Center
Anastasia's thesis is devoted to the creation of hierarchical sparse embeddings based on topic models.
Leading researcher of Antiplagiat company
Oleg's PhD thesis is built around fundamental problems related to the architecture of neural network models.
Search and Recommendations
Creation of hierarchical sparse embeddings based on topic models and more: we cross TM with transformer-based models. Using these embeddings to find and recommend articles.
Analysis of human behaviour
Psychotyping of a person based on data from a mouse and a touchpad. The task is based on the classification and clustering of semi-structured data.
Neural network topic modeling models and the application of TopicNet to ML tasks.
Neural Architecture Search
Automatic selection of deep learning models. Finding the optimal structure of a neural network for solving various types of problems. For example, classification and regression.
Application of models designed for working with texts to understand the language of genes. Training and analysing of gene models from the family of transformers for the restoration of gene networks of brain cells.
Development and implementation of algorithms for managing conversational skills, taking into account the goals of the user in the dialogue. Analysis of the dialogue structure, prediction of transitions between sub-dialogues.
The list of topics is updated with the arrival of new scientific leaders. Subscribe to the news - we will send you current topics of work and announcements of mentors. Send your resume when you find your area.
How to get to Science Club?
Attach your resume and motivation letter to your application. We want to know why you want to join us and what you did before.
Solve test task
The test task will be sent to the mail. You have 7 days to complete. Within a week after sending the result, we will write whether you have passed to the next stage.
Have an interview
We will talk with you at Zoom, get to know you and determine what tasks you are interested in. At the same time, we will try to find you a supervisor.
Meet the mentor
We will make an appointment with the leader you are interested in. You decide if you want to work together. If the mentor is not ready to start work, we will suggest choosing another one.
What work experience and knowledge is required to get to you?
- No work experience required! - Desire to engage in scientific activities - Student over 2 course - Programming experience in Python and knowledge of Git - Good English (Reading) - High average score in specialized subjects
Can I get in if I'm not from MIPT?
You can, there are no restrictions
What selection stages do I have to go through?
Fill out the form and send your resume
Solve a couple of test problems
Take a telephone and face-to-face intervie
How many hours per week should I devote?
A total of 20 hours, remote option.
Will I be paid?
Payment for research activities is not expected
I don't always have 20 free hours a week, can I take breaks from the research group?
Sure! You just need to warn in advance
Where can I ask questions if I cannot find the answer above?
You can ask any question about the internship by simply writing to us by mail firstname.lastname@example.org
Our students feedback
I was engaged in a task related to the analysis of banking transactions. Moreover, at our disposal were real data provided by the bank. It was nice to realize that what you were given as a task is really needed and in demand. Explore data, put forward hypotheses, propose solutions - there was a lot of freedom in the process of work, this is an opportunity to feel like a researcher.
MIL Team Researcher
After completing the Coursera specialization and mastering Python at an acceptable level, I wanted to get practice in working not only on test problems. Working in the MIL Team made it possible from the first week to take part in solving a real project, meet and communicate with professionals in their field, learn from the experience of a mentor (and visit SODA with her) and even go on a Yandex excursion from Vorontsov K.V. Thanks to the well-thought-out task structure, it was easy to follow the plan and get the final result in the form of a working model. After successful completion of the work, I was offered a job in the Laboratory. The past year has become a powerful boost in programming, the ability to create a convenient architecture and backend development.
Ex-Researcher at MIL Team
This is a great opportunity to move from learning tasks to real practice and research. Here I met a very friendly team of smart and interesting people who are always ready to help me. It is pleasant to apply the acquired knowledge to practical problems. I am very grateful to the laboratory for this opportunity.
Ex-researcher MIL Team
Fill in the fields