Edgar Parenti

AI Developer Tooling · Product Design & Architecture · Agentic Development

Summary

Product leader who designs, architects, and ships software products using AI coding agents as the engineering team. Manages full product development end-to-end (concept, spec, architecture, build, and ship) with Claude Code and Codex. Currently building the open infrastructure layer between agent connectivity and collective cognition: the governance, coordination, and shared memory that makes human-AI teams trustworthy, not just functional.

Spot the gap, design a solution, build it.

Experience

Technical Lead & Senior Content Architect Jul 2020 – Present
Red Hat · Portland, OR
Product Manager / Technical Product Owner Sep 2017 – May 2019
Datto · Portland, OR
Senior Product Manager, Networking Infrastructure Jul 2014 – Aug 2017
CDK Global (formerly ADP) · Portland, OR
Manager, Networks, Security & Unified Communications Jul 2012 – Jul 2014
Cambia Health Solutions · Portland, OR
Earlier Career Highlights (1995 – 2012)

Selected Projects

Jurati: MCP server providing shared team intelligence for agentic workflows. Git-backed knowledge repository with BM25 search, ambient coordination, signal classification, and privacy-aware context sharing across multi-agent sessions. 61 modules, 519+ tests. Python, Pydantic, MCP protocol.
EGON: Local control plane for human-agent software engineering. Orchestrates Claude Code and Codex with durable workflow state machines, agent handoff guardrails, collaboration rails for task/PR/verification tracking, and orphan process recovery. FastAPI, Pydantic, web shell UI.
Netballast: Configuration assurance system for automated networks. Formal state model with diagnostic vectors that classify drift by cross-layer agreement patterns. Lab-evaluated on 18-device multi-vendor estate: 4.6x triage reduction, blocked 8% of unsafe generated fixes. Grounded by formal academic paper. Python, FastAPI, React.
quatre-ulls: Gate doctrine and safety engine enforcing the Four-Eyes Principle in AI-assisted development. State machines for artifact progression, independent validator scoring, escalation policies. Pure-core architecture with pluggable runners and stores.

Skills & Technologies

Claude Code, MCP, AI agent development, Anthropic API, RAG, Python, FastAPI, React, Git, REST APIs, GitHub Actions, CI/CD, Docker, Ansible, OpenShift, Linux, PostgreSQL, Prometheus, Grafana, Product Management, Product Engineering, Agile/Scrum, Pragmatic Marketing

Education & Certifications