I build local evidence systems, semantic databases and LLM-assisted research tools.
My main project is The Ontology Machine: a local-first Windows system for turning document collections into evidence-bound SQLite corpus databases, then working with them through Query, Ontology and Taxonomy Agents.
The core idea is simple:
documents -> evidence artifacts -> semantic corpus DB -> ontology lenses -> computable knowledge work
The project grew from a personal research passion around theory-of-mind, functional primitives of the human psyche, document evidence, semantic control, and practical knowledge mining. Basically, the idea of why not try to condense thought? Since human thoughts are elusive, but code is not, the next obvious candidate is AI. And since AI is so damn fast and so good with semantics, and on the other hand so non-persistent in its output, why not capture its thoughts and turn them into something useful?
A local Windows application for:
- document ingestion
- source-grounded semantic search
- SQLite corpus databases
- evidence backlinks to page images
- Semantic Release based taxonomy/projection control
- ontology lens creation
- peer-review style knowledge mining
- local agent workflows
It is not a cloud app. It is not a thin chat wrapper. It is an attempt to make non-deterministic model work inspectable, evidence-bound and computable.
- human and machine co-reasoning
- evidence-bound AI systems
- local-first tools
- semantic control layers
- knowledge mining
- document intelligence
- digital memory architectures
- systems that expose uncertainty instead of hiding it
- The Ontology Machine repository: https://github.com/Sojemand/The-Ontology-Machine
- Latest release: https://github.com/Sojemand/The-Ontology-Machine/releases/latest
I am the author of Making Human Suspicion Computable, a position paper on evidence-bound knowledge mining, ontology lenses, and human-AI co-reasoning.
- Paper: Making Human Suspicion Computable
- ORCID: 0009-0002-1789-965X