Igor Kuznetsov

/
/
/
Hallo! I am a master's student in the Computational Modeling and Simulation programme at TU Dresden, focusing on aspects of differentiable simulation.

Broadly, I am interested in adaptive intelligent systems capable of operating robustly in a open-ended manner, in silico or embodied in the real-world. To progress in this direction, I study questions at intersection of reinforcement learning, simulation, and cognitive modeling.

I am a slow thinker and in my free time enjoy listening to Japanese ambient and wandering around new cities doing street photography.

Research

Safer Reinforcement Learning by Going Off-policy: a Benchmark

Igor Kuznetsov

NextGenAISafety Workshop @ ICML 2024

Paper | GitHub

Guided Exploration in Reinforcement Learning via Monte Carlo Critic Optimization

Igor Kuznetsov

AAMAS 2024

Paper | GitHub

Solving Continuous Control with Episodic Memory

Igor Kuznetsov, Andrey Filchenkov

IJCAI 2021

Paper | GitHub

Conditioning of Reinforcement Learning Agents and its Policy Regularization Application

Arip Asadulaev, Igor Kuznetsov, Gideon Stein, Andrey Filchenkov

BIG Workshop @ ICML 2020

Paper | IEEE Access

Research and Development Experience

2023-2025
Oxa (Oxford, UK), Software Engineer

Reinforcement Learning for self-driving cars safety

2020-2023
Giant.AI (Campbell, CA, remote), Member of Technical Staff

Reinforcement Learning for humanoid robots [robotics startup]

2020
JetBrains Research (St. Petersburg, Russia), Intern

Reinforcement Learning for complex locomotion in simulation

Education

2025 - Present
TU Dresden, Germany

MSc in Computational Modeling and Simulation

  ⊹ Basic Numeric Methods
  ⊹ Parallel Programming and HPC
  ⊹ Design Patterns and Frameworks
  ⊹ Good Scientific Practice

2017 - 2021
ITMO University, St. Petersburg, Russia

BSc in Computer Science and Computing Hardware

Summer Schools 🎒

2025

Tools

Python, C++, Vim, Git

Languages

Fluent
English
Beginner
German
Native
Russian

Blog

Updates

Oct 2025
Started MSc in Computational Modeling and Simulation at TU Dresden
Jul 2025
Accepted at EEML, will be visiting Sarajevo in Aug