Artificial Intelligence Research

Exploring general-purpose intelligence through multimodal models, reasoning systems, and embodied agents.

Our Vision

Yazhvin’s AI research aims to build systems that can understand, reason, and act across diverse domains — from science and creativity to language and robotics. We believe in general-purpose models that are interpretable, robust, and aligned with human goals.

Our work spans multimodal learning, neuro-symbolic reasoning, agent-based architectures, and ethical alignment. We're building next-generation foundational models and researching how these can be safely deployed in the real world.

Research Areas

Multimodal Models

Vision-language-action models that learn across text, image, video, and 3D environments.

Neuro-symbolic Systems

Combining deep learning with logic and planning to enable structured reasoning.

Autonomous Agents

Training embodied systems that can plan, act, and learn in simulation and the real world.

Language Understanding

From chat agents to reasoning assistants — we build and align large-scale language models.

Creativity & Code

Tools for AI-assisted design, scientific discovery, and generative programming.

Robustness & Safety

Ensuring models behave reliably, interpretably, and without bias or manipulation.

Open Research. Shared Intelligence.

We publish papers, release models, and support collaborations across the global AI community.