Alexey Skrynnik

PhD, Research Scientist at AIRI

profile-image.jpg

Moscow, Russia

As a Research Scientist with a PhD in Computer Science, my expertise centers on Artificial Intelligence and Machine Learning, particularly in the realms of applied Reinforcement Learning (RL) and Multi-Agent Systems. My work includes developing advanced RL algorithms and exploring the synergy between Planning and Learning. Notably, I’ve developed several state-of-the-art methods for decentralized multi-agent pathfinding, including Follower, MATS-LP, Switcher, and the POGEMA environment for evaluating these methods.

My contributions to hierarchical RL, particularly within embodied environments such as Minecraft, were highlighted by the ForgER approach that secured first place in the NeurIPS 2019 MineRL Diamond competition.

Furthermore, I’ve been leading efforts to combine Natural Language Processing (NLP) with RL to improve language-driven task solving, highlighted by my role in directing the RL track of the IGLU competition at NeurIPS 2021/2022.

News

Jul 27, 2025 Excited to have our paper “CrafText Benchmark: Advancing Instruction Following in Complex Multimodal Open-Ended World” in the Main Track at ACL 2025 in Vienna! We introduce CrafText, a new benchmark and environment for multimodal RL, designed to tackle instruction following in complex, dynamic settings. The paper also proposes several strong baselines to drive progress in this area. Here are the links to the paper
Apr 18, 2025 Excited to share that our paper, IDAT: A Multi-Modal Dataset and Toolkit for Building and Evaluating Interactive Task-Solving Agents, has been accepted to SIGIR!
It’s a great conclusion to the series of IGLU competitions. Read it on arXiv
Jan 23, 2025 I’m happy to share that our paper, POGEMA: A Benchmark Platform for Cooperative Multi-Agent Pathfinding, has been accepted to the ICLR-2025 Conference! Here are the links to the preprint on arXiv and the openreview.
Dec 10, 2024 I’m happy to announce that our paper, MAPF-GPT: Imitation Learning for Multi-Agent Pathfinding at Scale, has been accepted to the AAAI 2025 Conference! Here are the links to the preprint on arXiv and the open-source code.
Sep 05, 2024 I’m excited to announce our recent preprint titled MAPF-GPT, a GPT-like model designed for MAPF problems. It is trained using pure imitation learning on trajectories generated by LaCAM. MAPF-GPT performs exceptionally well on unseen instances and outperforms state-of-the-art learnable solvers such as SCRIMP and DCC. Here are the links to the preprint on arXiv and the open-source code.

Selected publications

  1. MAPF-GPT: Imitation learning for multi-agent pathfinding at scale
    Anton Andreychuk ,  Konstantin Yakovlev ,  Aleksandr Panov , and 1 more author
    In Proceedings of the AAAI Conference on Artificial Intelligence , 2025
  2. Idat: A multi-modal dataset and toolkit for building and evaluating interactive task-solving agents
    Shrestha Mohanty ,  Negar Arabzadeh ,  Andrea Tupini , and 5 more authors
    In Proceedings of the 48th International ACM SIGIR Conference on Research and Development in Information Retrieval , 2025
  3. Craftext benchmark: Advancing instruction following in complex multimodal open-ended world
    Zoya Volovikova ,  Gregory Gorbov ,  Petr Kuderov , and 2 more authors
    In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , 2025
  4. Gradual Optimization Learning for Conformational Energy Minimization
    Artem Tsypin ,  Leonid Anatolievich Ugadiarov ,  Kuzma Khrabrov , and 7 more authors
    In The Twelfth International Conference on Learning Representations , 2024
  5. Learn to follow: Decentralized lifelong multi-agent pathfinding via planning and learning
    Alexey Skrynnik ,  Anton Andreychuk ,  Maria Nesterova , and 2 more authors
    In Proceedings of the AAAI Conference on Artificial Intelligence , 2024
  6. Decentralized Monte Carlo Tree Search for Partially Observable Multi-Agent Pathfinding
    Alexey Skrynnik ,  Anton Andreychuk ,  Konstantin Yakovlev , and 1 more author
    In Proceedings of the AAAI Conference on Artificial Intelligence , 2024
  7. When to Switch: Planning and Learning for Partially Observable Multi-Agent Pathfinding
    Alexey Skrynnik ,  Anton Andreychuk ,  Konstantin Yakovlev , and 1 more author
    IEEE Transactions on Neural Networks and Learning Systems, 2023
  8. Interactive Grounded Language Understanding in a Collaborative Environment: Retrospective on Iglu 2022 Competition
    Julia Kiseleva ,  Alexey Skrynnik ,  Artem Zholus , and 8 more authors
    In NeurIPS 2022 Competition Track , 2023
  9. Pathfinding in stochastic environments: learning vs planning
    Alexey Skrynnik ,  Anton Andreychuk ,  Konstantin Yakovlev , and 1 more author
    PeerJ Computer Science, 2022
  10. Interactive grounded language understanding in a collaborative environment: Iglu 2021
    Julia Kiseleva ,  Ziming Li ,  Mohammad Aliannejadi , and 8 more authors
    In NeurIPS 2021 Competitions and Demonstrations Track , 2022
  11. Hybrid policy learning for multi-agent pathfinding
    Alexey Skrynnik ,  Alexandra Yakovleva ,  Vasilii Davydov , and 2 more authors
    IEEE Access, 2021
  12. Forgetful experience replay in hierarchical reinforcement learning from expert demonstrations
    Alexey Skrynnik ,  Aleksey Staroverov ,  Ermek Aitygulov , and 3 more authors
    Knowledge-Based Systems, 2021
  13. Hierarchical deep q-network from imperfect demonstrations in minecraft
    Alexey Skrynnik ,  Aleksey Staroverov ,  Ermek Aitygulov , and 3 more authors
    Cognitive Systems Research, 2021
  14. Camar: Continuous actions multi-agent routing
    Artem Pshenitsyn ,  Aleksandr Panov ,  and  Alexey Skrynnik
    arXiv preprint arXiv:2508.12845, 2025
  15. Advancing Learnable Multi-Agent Pathfinding Solvers with Active Fine-Tuning
    Anton Andreychuk ,  Konstantin Yakovlev ,  Aleksandr Panov , and 1 more author
    arXiv preprint arXiv:2506.23793, 2025