AI leadership
AI that makes engineering teams clearer and faster, not noisier.
I lead engineering organizations in using AI where it removes friction. I use AI coding assistants myself every day, so I know how the tools work in practice.
How I approach it
- Start with where teams lose time: unclear requirements, slow planning, repeated manual work.
- Apply AI to those points in the workflow. Engineers review what AI produces and own what ships.
- Build internal tooling the team can run and maintain itself.
- Write the practice down as a standard, so it spreads beyond the people who invented it.
- Set guardrails early: read-only access where possible, care with personal data, and care with untrusted input.
Hands-on with the tools
I am a hands-on practitioner of AI coding assistants in day-to-day engineering and leadership work.
- Cursor
- Claude Code
- OpenAI Codex
What I have done
- At Deposco, wrote the standard practice for turning product requirements into epics and stories. It uses structured workflows and AI assistance, and I wrote it with principal engineers.
- At SkAI, as VP of Engineering since May 2026, I am introducing AI-assisted software delivery and internal AI tooling for engineers.
- Built a shared library of AI-agent skills and routing rules. It gives engineers and AI coding agents guided, repeatable workflows for code audit, release readiness, ticketing, pull request review and test triage.
- Set quality gates for that library: validation, audits and a ticket-first pull request convention.
- Created read-only skills that let engineers and AI agents query internal operations tools with safety limits. They replace ad-hoc manual lookups for support and triage.
- Led a Slack-connected AI operations agent with scheduled monitors. They watch vendor status, release health and alarms, and post digests to the right teams.
- Put guardrails on AI agents: read-only database and cloud skills that block writes, personal-data minimization, hardening against untrusted input, and agent permission settings.
- Audited how AI agents use context, and compressed agent documents to reduce cost and improve reliability.
- Ran structured design-review sessions with product and engineering leads, then turned the outcomes into ticket-ready proposals. Rollout happens through reviewed tickets.
My AI work is internal: it supports how engineers plan, build and operate. It is not a customer-facing feature.
The foundation underneath
AI helps most where engineering basics are sound. I bring clear ownership, careful architecture and design review, and steady coaching of engineers and team leads. See engineering leadership and the full experience.