Alexei Korol · Applied AI Engineer
A multimodal agent benchmarked at 90% on GAIA Level 1, a web app live in production, and open-source agent tooling. Every claim below links to something you can run or read.
Flagship · cited portfolio RAG
Answers are constrained to projects, engineering notes, and the resume. Every supported claim links to its source.
Projects as evidence
A deployed multimodal agent that scored 90% (18 of 20) on the GAIA Level 1 final assignment of the Hugging Face AI Agents course. The course pass bar is 30%.
Student roommate-matching web app, live at app.roomeefy.pl since July 27, 2026. All production engineering is mine: auth, database, realtime, CI, and the guards that came out of launch day.
A repo-to-LLM-context tool: clone any repository and flatten it into a file-tree map plus token-aware code artifacts. Ships as a CLI, a FastAPI service, a React dashboard, and an MCP server.
Health and mana orbs for a game UI ("Vessels of Life & Mana"). One fragment shader draws both liquids, six AI-generated art plates provide the vessels, and the whole thing ships as a single self-contained HTML file.
Engineering notes
Build-time Markdown, route-level metadata, RSS, per-post social cards, code, numbers, and the wrong turn. No generic AI explainers.
Read lab notesI added BM25 to a dense RAG retriever, watched recall@1 regress, and used a cross-encoder to turn a wider candidate pool into a measurable win.
A single judge call hid variance. Three runs per rubric cost more, but exposed disagreement and created a concrete human-review gate.
The model call is optional; retrieval, citations, refusal checks, and a useful degraded answer continue when quota or configuration fails.