AI / ML · 2025 · Product Owner & Architect
Neurosymbolic AI System (KAG)
Architected and evaluated a knowledge-augmented retrieval pipeline for data-privacy compliance, combining ontologies with LLM retrieval.
A neurosymbolic system that grounds LLM retrieval in a formal ontology to reason over data-privacy regulatory documents — pairing the recall of retrieval-augmented generation with the precision of symbolic knowledge.
- Designed the knowledge-augmented retrieval (KAG) pipeline and its ontology layer.
- Built the evaluation framework used to validate accuracy on compliance questions.
- Led team coordination across architecture, validation, and delivery.
Knowledge graphs RAG Neo4j Evaluation