Tyler T. Procko, Ph.D. | AI Engineer | Florida, USA
procko.pro . linkedin . scholar . orcid . substack
I build systems where LLMs reason over structured knowledge — not just retrieve text, but traverse ontologies, respect formal semantics, and produce traceable, explainable outputs.
What I work on:
- 🤖 Ontology-grounded RAG and agentic LLM pipelines (Claude, GPT, Ollama, LM Studio, open-source)
- 🕸️ BFO/CCO-compliant knowledge graphs for defense, aerospace, and cybersecurity
- 🔍 ML provenance tracking with PROV-O — making AI decisions auditable
- 📄 37 publications spanning KG+LLM synergy, explainable AI, and applied ontology
What I believe: LLMs get radically better when you give them something real to reason over. Ontologies aren't legacy tech — they're the missing layer between retrieval and reasoning.
- 🔐 Principal Knowledge Architect @ stealth cybersecurity startup — ontology design, NLP model selection, Claude/GPT integration
- 💡 AI Consultant — ontology engineering, LLM evaluation (RLHF), knowledge graph construction
- 🌐 Building
procko.pro— my portfolio, generated entirely from a single YAML source file
- ⚔️
dissenter— multi-LLM adversarial critique framework for stress-testing AI outputs - 🔬
ProvTracer— dissertation tool: automated software traceability via KG + LLM synergy - ⭐
awesome-bfo— curated list for the Basic Formal Ontology ecosystem - 🎨
StableDiffusionEndToEndGuide— end-to-end SD guide (46+ stars) - 🦉
machine-learning-ontologies— OWL ontologies for ML concepts
