PINN
Public repository
Physics-Informed Neural Networks: research experiments, benchmarks, and comparative tests of PINN variants.
OpenML Engineer with 3 years of practical experience in experiments, model training, and production services across NLP, CV, and SciML.
I work at the intersection of applied research and ML engineering: from data pipelines and model development to fine-tuning, inference, and monitoring. My focus is NLP, CV, and LLM systems, including agentic RAG architectures for documentation-heavy domains. I turn research ideas into robust practical solutions with reproducible experiments, clear quality metrics, and production reliability. I also keep building expertise in diffusion and multimodal models, as well as MLOps practices.
Selected projects focused on applied ML engineering and research.
Public repository
Physics-Informed Neural Networks: research experiments, benchmarks, and comparative tests of PINN variants.
OpenPrivate repository
RAG system for technical documentation: indexing, retrieval, response generation, and chat history. Built with LangChain, semantic routing, agentic search, and MCP.
PrivatePrivate repository
Research repository for a [paper](https://ieeexplore.ieee.org/document/11314568/) on diffusion LLMs, focused on generation steps, resampling strategies, and quality-impact factors.
PrivatePublic repository
Platform for fine-tuning experiments and CV model comparison, built during the Leaders of Digital Transformation hackathon.
Open