Our Open Source Projects
#Rapid MVP Mindset implemented on our own projects
#1 - EdukaAI Studio
February 2026 - April 2026
Fine tune small LLM on your Apple Mac (M1, M2, M3). Part of EdukaAi Project.
Started for learning purposes and experimenting with fine tuning but it turns out as pretty solid project.
Brings me first hand experience how to fine tune small llm locally and importance of meta data in the process.
Build with AI and almost finished with AI. Except minor manual interventions AI "completed" this project .
Status: FAILED
#2 - AI Curator (ex EdukaAI curator)
March 2026 - April 2026
Dataset Management for LLM Fine-Tuning.
Import from files or live capture, organize samples, manage quality, and export for fine-tuning. Keep the track of meta data for each sample. The second part of EdukaAI project.
Written By AI, instructed by human, but without AI in the app.
Not finished. Reached the point where pretty capable model wasn't able to complete relatively simple task even with clear and direct instructions.
AI Curator on GitHub | AI Curator webpage
Lessons learned: Importance of tracking meta data for each sample.
Status: FAILED
#3 - Working Mind
April 2026 - June 2026
Non coding terminal AI agent build for researches who cares about privacy and security with persistent local knowledge graph.
This is the project that summarize all insights and knowledge I gained with previous work. I wanted to build simple terminal non coding AI agent that will be secure and private extensible by MCP servers.
Lessons learned: Once you create a knowledge graph you can export it as a dataset and fine tune small model who will provide answers from graph like a champ.
To read more about the idea behind visit docs page.
WorkingMind on GitHub | WorkingMind webpage
Status: FAILED