About Jen

AI systems designer, speaker, educator. I design the language system of AI and data products.

Jen Kelleman

I thrive in 0→1 environments, designing the human layer of agentic AI systems — the language, workflows, trust signals, and agent architectures that make autonomous AI usable, trustworthy, and learnable at scale.

Expertise

AI & Agent Systems

Agent Orchestration Model Context Protocol (MCP) Prompt Engineering Agentic UX Design LLM Integration Conversational AI Design Trust Signal Frameworks AI Evaluation Criteria MCP Servers Vibe Coding Workflows RAG Systems Responsible AI Human-in-the-Loop Design Multi-Agent Systems Semantic Layer Design

Content & Design Systems

Content Systems & Design Terminology Frameworks Information Architecture Design-to-Code Documentation Cross-Surface Audits Enterprise Style Guides SWIFT Terminology Studies Persona-Based Segmentation UX Writing Content Strategy Taxonomy & Ontology Accessibility (WCAG) Voice & Tone Systems

Engineering & Tools

Figma API Automation REST APIs Fluent UI Front-End Web Development Machine Learning SaaS Ecommerce Python Git & GitHub Hugo PowerShell

Leadership & Community

Public Speaking Community Building Mentoring & Teaching Cross-Functional Leadership Workshop Facilitation Thought Leadership Stakeholder Management

What I do

I name things for a living — column names, error messages, glossaries, metrics definitions — the words that sit between raw data and the people who need to use it. When a data product confuses someone, it’s almost never a model problem. It’s a naming problem. At Microsoft, I own that for Fabric Data Engineering. Over 15 years, I’ve shipped content design for autonomous AI features, led cross-stakeholder naming research through Technical Fellow and CVP approval gates, and built the terminology systems that AI-assisted features depend on.

Education & credentials

AI & Data

Carnegie Mellon University MS, Data Science (in progress)
Tufts University MS, Computer Science (in progress)
Microsoft Azure Data University Cloud Computing

UX & Communication

Emerson College MA, Writing and Publishing Digital Innovation
Deque University Designing an Accessible User Experience
Lafayette College BA, Psychology and English
GitHub Foundations certification badge GitHub Foundations
Microsoft Azure AI Fundamentals certification badge AI-900: Azure AI Fundamentals
Korn Ferry logo Korn Ferry: Leading the Future

What I believe

“Human language is the new UI layer.”

— Satya Nadella, Microsoft CEO

  • Most hallucinations trace back to the semantic layer. A column called amt gets grounded as “amount,” “amendment,” or “amortization.” Every one of those is a naming failure. Fix the name and you eliminate a class of failures no prompt engineering can touch.
  • Multi-agent systems fail at handoff boundaries. Two agents without a common ontology will break orchestration every time. Shared language is the fix.
  • The best systems build capability, not dependency. Scaffolded interactions that teach someone to build their own workflow outperform ones that just return an answer. Repeat-task rate drops. Task completion without re-prompting goes up.
  • Trust is a design problem. Users skip compliance disclosures. They look for confidence signals, source attribution, and uncertainty cues in the interface itself. If your AI can’t show its work, no governance doc will close that gap.

Let’s connect

Active in the Boston data and AI community through Boston PyLadies and Boston PyData. Always interested in speaking opportunities, teaching partnerships, and conversations about terminology governance, AI grounding, and semantic design.

Work with me → Book me to speak → See where I've spoken → LinkedIn