- 💼 Building multi-agent systems and GraphRAG pipelines as an AI Software Engineer at GlobalLogic
- 🎓 Studying Computer Science at Lviv Polytechnic National University
- 🏅 Microsoft Certified: Azure AI Engineer Associate – GenAI, NLP, Computer Vision
- 🧩 260+ problems solved on LeetCode (87 easy / 154 medium / 19 hard)
- 🔍 Currently digging into agent orchestration, retrieval quality and local inference
- 🚀 Previously founded Nevek – an OCR/CV tool that cut visual-content translation time by 75%
- 📄 Check out my detailed Resume / CV
Languages
AI & Machine Learning
Backend & Data
Databases
Cloud & DevOps
Tools
| Project | Description | Stack |
|---|---|---|
| citegraph-copilot | GraphRAG assistant over an 800-paper arXiv citation graph. Three agents over A2A, tools over MCP, human approval before the graph is mutated | LangGraph · Neo4j · A2A · MCP |
| retrieval-arena | Retrieval benchmark that runs in CI without an API key: IR metrics, bootstrap intervals and significance tests | Python · BEIR · sentence-transformers |
| local-docs-assistant | Fully offline RAG over your own documents, with a benchmark comparing local models on the same corpus | Ollama · Chroma · FastAPI |
| ai-inbox-triage | Ticket triage that routes what it is confident about and asks a human when it is not | n8n · FastAPI · LLM tool use |
| grid-outage-planner | Power grid outage scheduler: CIM XML topology, load forecasting and a priority queue over affected customers | Python · Gradio · NetworkX |
| cocktail-rag-chatbot | RAG chatbot answering from a vector index of 425 recipes, with retrieved sources shown in the UI | FastAPI · LangChain · FAISS |