Highlights
- An RNN memory model tracks pace, retention and misconceptions.
- A deep RL teacher agent adjusts topic order, difficulty and pacing.
- A RAG pipeline on Arch-Router-1.5B generates grounded explanations and practice problems.
- Also Top 23 and a Participation Prize at Student HackPad 2025.
Built with
RNN memoryDeep RLRAGPythonPyTorchReactNode.jsTailwind CSSsentence-transformers
Next project
NeuroFlowReal-time cognitive biofeedback