Principal Software Engineer in Test, leading AI adoption and building developer platforms at scale. I serve on the engineering AI leadership group at Instructure, where I've helped established the vision and roadmap for AI-augmented development across all phases of the SDLC.

• AI strategy & leadership — Shaping company-wide AI vision, standards, and implementation. Founded a monthly AI roundtable for cross-team knowledge sharing and maturation.
• AI skills & agents — Built AI agents that improve documentation, code testability, and test coverage with opinionated guardrails. Reduced project cycle time by 30%.
• AI-powered performance testing — Created skills and agents that cut onboarding time from 3–6 weeks to half a day — reclaiming 42+ engineering weeks of capacity.
• Claude Code plugins – Building a plugin ecosystem for Claude Code — skills, agents, and hooks that enforce opinionated workflows, TDD practices, and agentic automation patterns on top of the Claude Code CLI.
• Heavy agentic development — Quarantine is a real-world example of AI-driven development at scale, built almost entirely through agentic workflows.

Things I've learned while practicing agentic development. Some of these aren't new — they're engineering fundamentals that became more valuable once AI entered the picture.

• Specify what you want, verify what you got — Define intent precisely upfront and always validate outputs against that intent.
• Human readable, agent consumable — Design artifacts that humans can understand and agents can reliably parse and act on.
• Guardrails over guidelines — Use hooks, linters, and constraints to make the wrong behavior hard, not just discouraged. If it can be scripted, it should be.
• Measure what matters — Track outcomes like defect escape rate, MTTR, and lead time — not just how fast the code ships.
• AI accelerates what already exists — Healthy engineering practices scale faster with AI, but so does dysfunction; quality foundations aren't optional.