Skip to content
View alexkorol's full-sized avatar
15 followers · 3 following

Achievements

Achievement: Pair ExtraordinaireAchievement: Pull Sharkx3Achievement: YOLOAchievement: Quickdraw

Achievements

Achievement: Pair ExtraordinaireAchievement: Pull Sharkx3Achievement: YOLOAchievement: Quickdraw

Block or report alexkorol

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
alexkorol/README.md

Hi, I'm Alex Korol

LinkedIn Portfolio Email

AI engineer focused on LLM agents, evaluation, and RAG. Hugging Face Agents certified and Azure AI certified. Seattle.

GAIA benchmark agent

Passed the Hugging Face AI Agents course (July 2026) with a deployed multimodal agent that scored 90% (18/20) on the GAIA Level 1 final assignment; the course pass bar is 30%. Hybrid smolagents architecture (CodeAgent with a ToolCallingAgent fallback), web search, revision-aware Wikipedia retrieval via the MediaWiki API, and handling for spreadsheets, PDFs, audio, video, and images. Live on Hugging Face Spaces (certificate).

Tools

  • repo2GPT - turns any repository into an LLM-ready repo map plus consolidated code bundle. CLI, FastAPI service, React dashboard, and an MCP server.

Game UI and generative media

  • WIZARD - interactive game-UI experiments. Flagship: a WebGL2 health/mana orb engine where one fragment shader draws the liquids, built from six AI-generated art plates, shipped as a single self-contained HTML file (live demo linked in the repo).
  • pixel-perfecter - turns diffusion-generated "pixel art" into clean grid-aligned sprites: grid estimation, color-cell reconstruction, PySide6 GUI plus batch CLI.
  • delaford - heavily reworked fork of an open-source multiplayer RPG; hardened WebSocket authorization at the dispatcher boundary with unit tests that expect spoofed player references to be rejected.

Currently building a spaced-repetition learning app (React/TypeScript PWA with a native iOS build in progress).

Pinned Loading

  1. repo2GPT repo2GPT Public

    Clone any GitHub repo and flatten it into a file-tree diagram plus one consolidated code file - quick repo context for LLMs

    Python 10 2

  2. verdigris verdigris Public

    Verdigris — a WASD-first multiplayer ARPG.

    JavaScript 1

  3. WIZARD WIZARD Public

    AI-built toolbox of interactive game-UI experiments - WebGL2 ARPG orbs, PoE-style geometric skill tree, modular itemization engine (Vesselforge), permadeath RP meta-progression, falling-sand sandbo…

    HTML

  4. pixel-perfecter pixel-perfecter Public

    Turns diffusion-generated "pixel art" into clean grid-aligned sprites — grid estimation, color-cell reconstruction, PySide6 GUI + batch CLI

    Python