AIIR
MIPT · Master's Program

Artificial Intelligence and Interdisciplinary Research

Artificial intelligence for a complex world

A master's program at the Phystech School of Applied Mathematics and Informatics, MIPT — for future research group leaders, team leads, and architects of socially impactful IT systems.

Apply See student research
2 yrs
4 semesters, full-time
17+
courses, from MLOps to Roman law
16
faculty members
1
research seminar at the core of it all
The Program

Three foundations

Engineering foundation

The modern machine learning stack: from core ML to statistics, game theory, and MLOps.

Research seminar & thesis

Formulating research questions, academic conferences, and the master's thesis — the core of the program and its most challenging part.

Sources, law, texts

Rigorous required courses in ancient history and Roman law, and research at the intersection of machine learning and digital humanities.

Program partner — VTB Bank. The program is developed in partnership with the bank, keeping students connected to the professional world of industrial AI challenges.
Student Research · with the MASKI Laboratory

Research beyond the classroom

The program's research life runs jointly with the MASKI Laboratory at the Phystech School of Applied Mathematics and Informatics — a research lab in its own right, distinct from the department that hosts the program. Its research seminar is where students present real studies, from Latin word embeddings to the syntax of Imperial Russian court rulings. Research lead: Ivan Selivanov.

Latin corpora & NLP Imperial Russian legal records Historical data engineering
Latin corpora & NLP

Static vs. contextual word embeddings for semantic search in classical Latin

Which type of embedding captures meaning better across a corpus of classical Latin — a head-to-head comparison on the semantic-search task.

I. Semenov · seminar note
Latin corpora & NLP

LLM-based semantic similarity for detecting intertextuality in classical Latin

Six current Latin and multilingual embedding models, benchmarked against a gold-standard corpus of annotated intertextual parallels from Valerius Flaccus' Argonautica.

I. Yakovenko · seminar note
Latin corpora & NLP

Interpretable authorship attribution for Latin texts of the 3rd–9th centuries

A comparison of attribution methods with one hard constraint: every result must be reliably interpretable, not just accurate.

E. Tolchenitsyna · seminar note
Imperial Russian legal records

Teaching OCR to read pre-1918 Russian

A tuned ABBYY FineReader pipeline vs. LLM-based OCR on rulings of the Governing Senate — Imperial Russia's supreme court — printed in pre-1918 spelling. Which reads the old script better?

E. Smagina · seminar note
Imperial Russian legal records

Semantic analysis and visualization of Senate decisions

Do modern embedding models uncover stable semantic clusters in the Senate's appellate rulings?

E. Lykova · seminar note
Imperial Russian legal records

Measuring the complexity of legal texts from a small sample

The syntactic complexity of Senate rulings, measured by maximum dependency length within a sentence.

M. Abdulloeva · seminar note
Historical data engineering

Turning semi-structured historical biographical records into machine-readable data

Deterministic vs. non-deterministic extraction pipelines, applied to a biographical database of Imperial Russia's service elite.

K. Morozov · seminar note
MASKI Laboratory · Russian Science Foundation grant

Silius Italicus' Punica: translation, historical realia, reception

The laboratory's flagship philological project: the longest surviving Roman epic — an erudite Flavian poem on the Second Punic War, rich in geographic, ethnographic, and historical detail — still has no complete Russian translation. Supported by the Russian Science Foundation.

The research seminar is open — to give a talk, write to Ivan Selivanov at selivanov.is@mipt.ru.

Faculty

Who teaches here

Mathematicians, ML engineers, historians, and philologists — from MIPT, MSU, HSE, the Russian Academy of Sciences, Yandex, and Kaspersky Lab.

Ivan Selivanov

Program director · Research lead

Research fellow at the MASKI Laboratory, lecturer at the Department of Machine Learning and Digital Humanities, head of the program

Konstantin Vorontsov

Academic director

Dr. Sci. in Physics and Mathematics, head of the Department of Machine Learning and Digital Humanities, MIPT

Radoslav Neychev

Machine learning

PhD in Physics and Mathematics, head of an ML development group at Yandex, deputy head of the department

Vladislav Goncharenko

MLOps

Founder of girafe.ai, senior lecturer at Lomonosov Moscow State University and at MIPT

Dmitry Zaytsev

Ancient history · Greek

PhD in History, research fellow at the MASKI Laboratory, associate professor at MIPT

Anna Kulashova

Leadership · Strategy

Managing director at Kaspersky Lab, director of the Center for Game Design and Development Education at MIPT

Ilya Petrov

Research seminar

PhD in Engineering, research fellow at the International Laboratory for Applied Network Research (HSE) and the Institute of Control Sciences RAS

Alexey Savvateev

Game theory

Dr. Sci. in Physics and Mathematics, corresponding member of the Russian Academy of Sciences, professor at MIPT

Dmitry Gubanov

Research seminar

Dr. Sci. in Engineering, head of the Aizerman Laboratory of Choice Theory and Decision Analysis (ICS RAS), professor at MIPT

Maxim Beketov

Statistics · Time series

Research fellow at the International Laboratory of Stochastic Algorithms and High-Dimensional Data Analysis (HSE), senior lecturer at MIPT

Yaroslav Kotov

Industry practice

PhD student at the department, lead specialist at the AI Platform Solutions Center, OTP Bank

Alexey Nikitin

Game theory

PhD in Physics and Mathematics, associate professor at Lomonosov Moscow State University, lecturer at HSE

Alexander Tonis

Game theory

PhD in Economics, associate professor at the New Economic School and at MIPT

Nina Spichenko

Roman law

PhD in History, associate professor at the State Academic University for the Humanities and at MIPT

Ekaterina Vikhrova

English for AI

PhD in Philology, associate professor at the Department of Foreign Languages, MIPT

Tikhon Davydov

Latin

PhD in Philology, senior research fellow at the MASKI Laboratory, senior lecturer in classical philology at Lomonosov Moscow State University

Curriculum

Two years, semester by semester

Semester 1 autumn
Research seminar: contemporary problems of applied mathematics and informaticsGubanov · Petrov · Selivanov
Introduces students to the work of the department and the MASKI Laboratory: semantic search across Latin corpora, OCR for pre-1918 Russian texts, computational analysis of Imperial court rulings, and other digital-humanities problems.
Probability theory and foundations of mathematical statisticsBeketov
An essential foundation for the further study of AI methods.
Foundations of machine learningNeychev
Covers both classical approaches and the foundations of deep learning.
Software developmentstaff
Programming fundamentals in Python, core OOP concepts, advanced language structures, and the developer tools needed for complex projects.
LeadershipKulashova
Applying leadership concepts and styles in context, management skills, teamwork, identifying target audiences, and communication with teams and stakeholders.
Semester 2 spring
Applied statistics and data analysisBeketov
Advanced topics in statistics and methods for solving applied data-analysis problems.
Applied machine learning methodsstaff
Key results achieved in natural language processing, reinforcement learning, and computer vision.
MLOps: deployment and maintenance of ML solutionsGoncharenko
Storing, processing, and using large datasets in model training; serialization and production deployment. A bridge between ML theory and software practice.
History of antiquityZaytsev
The Greco-Roman world: institutions, society, and events, with an emphasis on source criticism — what evidence our knowledge of antiquity actually rests on, how to work with incomplete information, and how to tell fact from interpretation and reconstruction.
Strategic managementKulashova
Developing and implementing organizational strategy: environment analysis, decision synthesis (Ansoff, Blue Ocean, SAFe), balanced scorecards, agile approaches, change management.
Semester 3 autumn
Game theorySavvateev · Nikitin · Tonis
Fundamental mathematics and game theory: a two-part course on mathematical foundations and strategic interaction.
Time series analysisBeketov
Forecasting, anomaly detection, and pattern recognition in sequential data — from classical ARIMA models to modern ML methods used in finance, economics, meteorology, and marketing.
Foundations of Roman lawSpichenko
How Rome built a legal system out of concrete disputes: foundational legal principles and case studies showing how an abstract rule starts to work in a real-world situation.
Semester 4 spring
Master's thesisSelivanov · Gubanov · Petrov
The culmination of the research track: an independent study at the intersection of machine learning and the humanities, defended before the department.
English: modern artificial intelligenceSemesters 1–2 · Vikhrova
Professional communication in AI: evolution of the field, deepfakes, superintelligence, science and art, ethics of AI adoption. Seminars, group projects, and work with English-language materials.
Latin (elective)Semesters 1–2 · Davydov
A language with a rich grammatical system that trains memory and precision of thought — and a gateway to ancient history, Roman law, and the roots of modern European languages.
Ancient Greek (elective)Semesters 1–2 · Zaytsev
Read ancient texts in the original: the fundamentals of grammar and translation, and, through the language, Greek culture and its intellectual tradition. Prior Latin helps but is not required.
Research workSemesters 2–4 · Selivanov · Gubanov · Petrov
Reading and analyzing papers, formulating research problems, writing a literature review and the master's thesis.

humanities track  ·  click a course to expand its description

Outcomes

What you will be able to do

Conduct research

Formulate research questions, work with the literature, present at conferences, and produce in-depth academic research.

Work with the modern ML stack

Apply machine learning, statistics, game theory, and MLOps to academic and industrial problems.

Navigate complex domains

See the world more broadly than the model does.

Admissions

Your research starts with an email

Admission begins with a conversation. Write to Ivan Selivanov, the program director, at selivanov.is@mipt.ru — tell us about your background and what you'd like to work on.

Questions about the program are welcome too. There's no wrong first email.

Email the program director