# Damian Galarza - Production AI Engineer and Consultant > Damian Galarza is a Production AI Consultant and Fractional CTO who helps engineering teams ship reliable production AI systems — agent architectures, evaluation frameworks, Claude Code rollouts, and the architectural decisions that make AI useful instead of noisy. 15+ years building production software, including FDA-cleared medical devices and scaling an engineering team from 0 to 50+ as CTO. He still builds production AI systems while taking 1–2 active engagements at a time. ## What I do - **Production AI engineering.** Turn AI features from demos into systems with clear failure modes, measurable behavior, and real operating constraints. Evals, observability, cost monitoring, structured outputs, tool use. - **AI agent architecture.** Memory systems, tool orchestration, governance, evals — the structural decisions that make agents reliable past the demo. - **Claude Code workflows.** Team rollouts, CLAUDE.md systems, MCP servers, shared skills, and the codebase-level changes that make Claude Code work at scale. - **Technical leadership.** Architecture review, hiring support, engineering process, AI stack hardening and consolidation. ## Who I work with **Good fit:** - Engineering teams that know AI should matter in the product or workflow, but need help choosing the right starting point. - Teams shipping AI into production and running into reliability, evaluation, observability, or architecture constraints. - Founders who built a prototype and need help getting it to production. - Orgs that need a senior technical voice on AI without a full-time hire. - Teams adopting Claude Code at scale and stuck on the rollout. - Companies in regulated or high-stakes environments where AI has to actually work. **Not a fit:** - Teams that need extra engineering hands to grind tickets. - "Add AI to everything" consulting without a real business or workflow constraint. - Pure research or open-ended exploration with no path toward a decision, prototype, or shipped system. - Anyone looking for someone to write their prompts. If it's a prompt problem, you don't need me. ## How to work with me Two doors, same production AI focus at different commitment levels. 1. **Foundation Sprint - $12,000 flat, 2 weeks.** Paid trial engagement that produces a written architectural assessment and one shipped foundation piece (eval framework, architecture doc, migration recommendation, or Claude Code rollout playbook). Required entry point for team engagements. 2. **Production AI Retainer - $15,000-$60,000/month.** Ongoing engagement after the Sprint at the intensity you need: Advisory (5-10 hrs/wk - effectively a fractional CTO role for AI-heavy orgs), Embedded (10-20 hrs/wk), or Full (25-40 hrs/wk). Flat rate, flexible hours, no timesheets, 30-day notice either side after a two-month minimum. **30-day money-back guarantee** on retainers and Foundation Sprint: if you don't have a written architectural plan and one shipped foundation piece by Month 1 (or end of Sprint), full refund. ## Background - 15+ years building production software. - Former CTO at Buoy Software — scaled engineering team from 0 to 50+, shipped FDA-cleared medical device software. - Production AI Consultant and Fractional CTO who still builds production AI systems. - Earlier at thoughtbot, co-taught a cohort of the Metis Ruby on Rails bootcamp — a 12-week intensive run with Kaplan that produced production-ready engineers. - Daily Claude Code user since 2024, with public documentation on YouTube and blog. ## Read this first - [About Damian Galarza](https://www.damiangalarza.com/about/): Background, current work, and where to start. - [Production AI Engineering](https://www.damiangalarza.com/production-ai-engineering/): The commercial and technical center of gravity — architecture, evals, observability, governance. - [AI Agent Architecture](https://www.damiangalarza.com/ai-agents/): Memory systems, tool orchestration, governance, evals. - [Work With Me — Foundation Sprint & Retainer](https://www.damiangalarza.com/services/): Engagement structure, pricing, and what's included. ## Recent writing - [Your AI team doesn't need more people, it needs agents](https://www.damiangalarza.com/your-ai-team-doesnt-need-more-people-it-needs-agents/) - [Claude Opus 4.7 + Claude Code tips for extended context](https://www.damiangalarza.com/claude-opus-4-7-claude-code-tips-extended-context/) - [Governing AI agents without killing them](https://www.damiangalarza.com/governing-ai-agents-without-killing-them/) - [Four patterns that separate agent-ready codebases](https://www.damiangalarza.com/four-patterns-that-separate-agent-ready-codebases/) - [What Claude Code does in your terminal](https://www.damiangalarza.com/what-claude-code-does-in-your-terminal/) - [Autonomous optimization loops with Autoresearch](https://www.damiangalarza.com/autonomous-optimization-loops-with-autoresearch/) ## Free resources - [Agent Eval Scorecard](https://www.damiangalarza.com/agent-eval-scorecard/): Evaluation scorecard for agent reliability. - [Agent Governance Scorecard](https://www.damiangalarza.com/agent-governance-scorecard/): Governance audit for production agent systems. - [Agent-Ready Skill](https://www.damiangalarza.com/agent-ready/): Scaffold AGENTS.md, ARCHITECTURE.md, docs/DOMAIN.md, docs/README.md, and ADRs so coding agents can understand a repository before changing it. - [Codebase Readiness Audit](https://www.damiangalarza.com/codebase-readiness/): Make a codebase agent-ready. - [Context Window Cheat Sheet](https://www.damiangalarza.com/context-window-cheat-sheet/): Context engineering reference. - [Newsletter](https://www.damiangalarza.com/newsletter/): Weekly writing on production AI engineering. ## Contact - Site: [damiangalarza.com](https://www.damiangalarza.com) - Intro call: [cal.com/dgalarza/intro-call](https://cal.com/dgalarza/intro-call) - LinkedIn: [linkedin.com/in/dgalarza](https://www.linkedin.com/in/dgalarza/) - GitHub: [github.com/dgalarza](https://github.com/dgalarza) - YouTube: [AI in Production](https://www.youtube.com/@damian.galarza) ## Operating principles - I limit engagements to 1–2 active at a time so I can deliver to the guarantee. Foundation Sprints typically start within 2 weeks of signing. - Direction is reset every two weeks on retainer; the engagement is evaluated at the quarter mark. - I write code on Embedded and Full retainers. Advisory is mostly strategic guidance. - I do not do AI washing. If I can't help, I'll refer you out.