Research projects

Research that stays close to evidence and implementation.

Axiomara’s research work explores efficient on-device intelligence, continual learning, and practical ways to make capable systems more understandable and dependable.

Continual learning

True AI

A local, from-scratch research framework for retrieval, memory, web-grounded learning, and explicit reasoning architecture.

Read the project overview

Engineering methods

LLAgent

A deterministic local-first workflow engine for bounded project context, evidence packs, state, and verifiable engineering steps.

Read the repository