Slava Fuodorov

I teach physics to eighth-graders at the Orthodox Gymnasium of St. Sergius of Radonezh in Akademgorodok, and I am writing an interactive textbook for grades 7–9 to go with it: every lesson is a notebook with theory, an experiment and a numerical simulation in Python.

I teach at Novosibirsk State University: lectures on Python in physics research for the Applied Mathematics and Physics programme, and seminars on the basics of artificial intelligence for the Physical Informatics track. I try to make students write scientific code that someone else can still open six months later.

I build accelerator control and beam simulation software at the Budker Institute of Nuclear Physics and at the SKIF synchrotron radiation facility. I wrote SCAUT and ACCUMULATOR — frameworks for automating experiments on the SKIF injector, with digital twins on Elegant, EPICS IOC and Docker — and SAVA, a conversational LLM assistant for operators. I compute beam envelopes and trajectories with relativistic difference schemes in Python and C++, and optimise them with genetic algorithms and Bayesian optimisation.

I built a platform that automates the research cycle for Text-to-SQL tasks: I adapted an AI agent (AI Scientist by Sakana AI) to generate hypotheses, run experiments, write code and analyse results, and added automatic report generation. A full research cycle came down to under $5.

I led a team of five developers on a fintech product built with Django, PostgreSQL, Celery, RabbitMQ and React. I built an ML service that assigns CPT medical codes automatically — microservices, serverless, AWS Lambda, DynamoDB, GraphQL — and an IntelliJ plugin for static analysis on top of SonarQube. I estimated work, ran code reviews and gave internal seminars.

I developed and tested an infrared horizon sensor for the Norby CubeSat. I wrote the test software and ran the trials whose results fed into the satellite launch in 2020.