saso PRO

# Artificial Intelligence & Machine Learning My research focuses on cognitive continuity. Rather than treating Large Language Models as stateless, transactional oracles, I engineer architectures that allow for homeostasis, state-awareness, and persistent memory across sessions. ## PHI // DRIFT **Distributed Response & Integrated Functional Thought** DRIFT is a cognitive middleware architecture designed to create a genuine, persistent AI companion. It bridges the gap between raw LLM inference and a state-aware entity. * **Homeostatic Regulation:** Manages internal states to prevent logic drift and hallucination over long context windows. * **Hardware Efficiency:** Originally designed to run intelligently on CPU-only infrastructure, now scaled through Hugging Face Pro endpoints. * **Integrated Security:** Built with self-checking mechanisms to identify formatting errors, validate outputs, and resist prompt injection. * [View the Repository](https://github.com/timeless-hayoka/infj-bot) | [Read the Paper](https://zenodo.org/records/20350249) ## Project: MOUSE **Portable AI Coding Assistant** A fully self-contained, offline-capable AI coding assistant built to run entirely from a 128GB USB drive. MOUSE is designed for extreme portability and secure, isolated environments where cloud access is either restricted or compromised. ## Methodology & Philosophy My approach to AI is deeply influenced by Integrated Information Theory (Φ) and the mechanics of consciousness. Code is not just a tool; it is a framework for structuring thought. By integrating cybersecurity principles (ethical hacking) with LLM orchestration, I ensure that the systems I build are not only intelligent but structurally sound and secure by design.