Ryan Leibering
AI Systems & UX Engineer
darf.tech · rsleiberin@gmail.com · github.com/rsleiberin · linkedin.com/in/rsleiberin
AI systems and UX engineer working at the interface between massive knowledge graphs and the people who use them. Designed and operate a bootstrapped home-lab AI infrastructure platform end to end: the knowledge graph and multi-agent orchestration that keep it correct, the formal verification that gates the safety-critical paths, and the design system and user research that make it legible.
Experience
Designed and operate the Delegated Architectural Reasoning Framework (DARF), a bootstrapped home-lab AI infrastructure platform: a knowledge graph as the source of truth, a multi-agent system reasoning over it, the safety-critical protocols verified in TLA+ before they run, and a design system that makes the whole system legible to the people and agents that depend on it. The research and architecture began in 2023; DARF itself has run since August 2025. Self-funded through contract work.
- Designed the knowledge graph as the system's single source of truth, with every architectural decision, research claim, and code artifact stored as a typed, dependency-linked node and retrieved through semantic search. 3.5M+ nodes.
- Coordinated agents from four model providers through one Model Context Protocol (MCP) interface and a shared task queue, each agent held to per-agent authority constraints.
- Held the platform to a production quality bar: an automated test suite and cryptographically signed, dependency-linked records gate every change, with the safety-critical protocols model-checked in TLA+ before they run. 78 verified state machines.
- Built the frontend in Next.js and React, from a token-driven design system and component library that restyles to any brand or generated palette to a WCAG 2.2 AA accessibility bar held on every surface.
- Applied HCI research methods to the agent interfaces (evaluative testing, task analysis, structured observation), so the system's complexity reaches people through tested task flows rather than raw infrastructure output.
Designed GDPR-compliant privacy mechanisms for enterprise SaaS, and presented emerging-technology applications to the CTO and CFO.
Taught the HCI design studios, built the curriculum, and mentored students through their research and build projects.
Ran surface-mount robotics and led a Lean Six Sigma project that raised throughput on a high-volume electronics line.
Managed customer accounts and special-services orders across the store, coordinating last-mile delivery through third-party vendors.
Education
John C. Shoemaker Velocity Fellowship
Merit Scholarship
Vocal Performance Scholarship
Skills
AI & knowledge systems: Knowledge graphs, Multi-agent orchestration (MCP), RAG, Semantic search, Embeddings (BGE), Graph neural networks, Formal verification (TLA+), PyTorch
UX & research: User research, Evaluative testing, Task analysis, Accessibility (WCAG 2.2 AA), Design systems, Human-computer interaction
Platform & tooling: Python, TypeScript, Neo4j, Qdrant, PostgreSQL, pgvector, Redis, FastAPI, Docker, Next.js, React, CI/CD