Graphlink (formerly Graphite) is a local-first, graph-based AI workspace for building branching reasoning workflows, running agent plugins, and orchestrating multiple AI providers in a visual environment designed for complex problem solving.

Register of deep learning-based products in radiotherapy

Sheep Herding Simulation is a realistic, modular, and AI-driven sheep herding simulation built entirely in Python using the Pygame library. This project delves into the fascinating world of emergent behaviors and artificial intelligence, showcasing advanced flocking algorithms for sheep and strategic AI for the shepherd.

🧠 Offline Ollama Inference Engine built with Java & Spring Boot. Run LLMs like Llama3, Mistral, Phi-3 Mini locally for privacy, multi-model switching, and reliable AI inference. Perfect for customizable, offline AI applications. 🤖

VidGenius is an AI-powered web app that extracts transcripts from YouTube videos, summarizes them using ChatGPT, and provides accurate translations with Google Gemini.

The Agent That Ensures Mastery. An adaptive learning system built with LangGraph and RAG that makes sure you truly understand. It focusses on concept based understanding and triggers a Feynman Remedial Loop to re-teach failed concepts using analogies.

Text Mining project - Artificial Intelligence, Vrije University,2025

An AI-powered “Classroom Co-Teacher” that enables UK schools and youth organisations to deliver high-quality AI and robotics projects with minimal preparation, even when teachers are not subject experts. The platform plans lessons, coaches pupils through hands-on projects, and tracks progress automatically, all aligned to the UK curriculum.

A retrieval-augmented generation VLM pipeline for Earth Observation disaster triage

AI-powered mental health chat assistant

CheckMe eliminates manual paper checking by using a flatbed scanner, AI-powered OCR (Google Gemini), and a mobile app to automate the entire grading workflow — from scanning answer keys to scoring student sheets and storing results in the cloud.

This assignment repository demonstrates the existance of deep double descent phenomena for the deep learning models usually encorporated with a clinical context. For this, a famous ECG dataset, PTBXL has been employed.