🤓 Latest Submissions
AI Knowledge Detective
Knowledge Detective is a next-generation AI enterprise knowledge graph engine built to solve the problem of organizational amnesia. Every day, valuable context gets buried across disconnected tools — GitHub commits, Gmail threads, Google Calendar meetings, and Slack conversations — and traditional AI search only makes it worse, since standard RAG retrieves isolated text chunks without understanding how people, documents, and code actually relate to one another. Knowledge Detective solves this with a dual-engine architecture that pairs Qdrant for fast semantic search with Neo4j for structured relationship mapping, so answers are grounded not just in what was said, but in who said it, when, and what it connected to. On top of this foundation, the platform adds dynamic identity resolution (unifying a person's scattered handles and emails into one canonical profile), multi-stage LLM verification to eliminate hallucinations, and gap detection that surfaces decisions proposed but never followed through. The result is a fully containerized, production-ready system — with working GitHub, Gmail, and Calendar connectors, a multi-agent FastAPI backend, and an interactive React dashboard featuring citation-backed chat, graph visualization, and timelines — deployable either fully on-premise on AMD Radeon and Instinct GPUs for maximum data sovereignty, or via managed cloud inference through Fireworks AI. It's positioned as the intelligence layer for enterprises that want the power of AI-driven institutional memory without sacrificing control over their IP.
13 Jul 2026
