Dan Mercede
AI Systems Architect - Governed Agent Runtimes, MCP, and AI Control Planes
San Diego, California, United States
Actions
Dan Mercede is a founder and AI systems architect focused on governed agent runtimes, MCP/tool governance, AI risk controls, retrieval systems, and production AI control planes.
His work centers on a practical production question: how do you let AI agents do useful work without giving them uncontrolled mutation authority? Dan builds execution environments where authority is evaluated at action time, ambiguous state fails closed, material actions produce auditable receipts, and behavior is constrained across time.
His public proof artifacts include mcp-context-budget, a local-first MCP context budget and tool-selection verifier for agentic coding environments; schemafit, a provider-aware structured-output and JSON-Schema CI linter for OpenAI, Anthropic, Gemini, Mistral, and Cohere; and failclosed, a Python merge-admission gate for AI-written code. His talks focus on MCP/tool governance, structured-output reliability, agent authority, adversarial code review, event-log patterns for agent state, and the boundary between advisory AI and enforceable controls.
Links
Area of Expertise
Topics
NODES 2026 Sessionize Event Upcoming
Dan Mercede
AI Systems Architect - Governed Agent Runtimes, MCP, and AI Control Planes
San Diego, California, United States
Links
Actions
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top