How I lead data
Most companies have more data than they can explain, and they are about to hand it to AI agents that will explain it confidently and wrong. The job of a data leader has changed: it is now deciding what context exists, who owns it, and which of it an agent may act on.
Strategy from the leadership seat, architecture from the platform seat, and enough hands on the keyboard to know whether the plan is real. That is the job as I understand it, and it is the job I do.
Principles
Govern context before you scale retrieval
Metric definitions, semantics, and lineage live in version control with owners. An agent that cannot point at where its definition came from is not ready for production.
Agents propose, people decide
Every automated workflow I have shipped keeps a human at the review gate. Speed came from removing patterned work, not from removing judgment.
Skills are the org's style guide
Conventions written down once, read by engineers and agents alike. Onboarding a person and onboarding an agent should be the same document.
One vocabulary across domains
A semantic layer is a leadership decision, not a tooling choice. Finance, product, sales, and support answer questions with the same words or they answer them wrong.
Grow people, keep them
Built a team from 2 to 8 with 100% retention and three promotions to senior. Retention is the metric that tells you whether the roadmap is honest.
Record
What the principles have looked like in practice.
- ZendeskLeads AI-readiness for a petabyte-scale platform used by 5+ engineering teams
- ZendeskSet engineering best practices adopted across teams; mentors engineers on multi-system, long-term projects
- AchieveDrove data-driven roadmaps with data scientists, analysts, product managers, and business leaders
- Noodle.aiBuilt and managed a data engineering team from 2 to 8; hiring, performance reviews, career guidance
- CognizantPlanned and delivered multi-team, long-term programs with external dependencies across three countries
- CommunityKeynote and conference speaker on AI-ready data platforms; hackathon judge; mentor to engineers in the US and Singapore