How I run engineering, written down so you know what you’re getting before our first conversation.
A vague goal is a risk. My first job is to turn it into a plan a team can start on Monday.
I start from the business outcome, name the unknowns, and break the work into slices that each prove something. The first slice is the riskiest assumption, not the easiest screen.
Signals I watch
Decisions outlive the meeting. Put the options, the choice and the tradeoff on paper.
Significant choices get a short decision record: context, the options considered, what we chose, and what we are giving up. It keeps debates short, onboarding fast, and reversals honest.
Signals I watch
Debt is a financing decision. Take it on deliberately, record it, and pay it down where it costs interest.
I budget capacity for paying down debt every cycle and spend it where debt slows the roadmap: hot paths, flaky tests, risky deploys. Cosmetic debt waits.
Signals I watch
Small stacked PRs, continuous integration, and releases so frequent they stop being events.
Batch size is the lever. Small changes get reviewed properly, deploy safely and roll back cheaply. I invest in the pipeline before asking a team to go faster.
Signals I watch
Surprises are cheapest in week one. I pull integration, data and security risk to the front.
Walking skeletons end to end before features, real data before demos, and security and accessibility as release gates rather than launch-week audits.
Signals I watch
Maintainable beats clever. Boring technology, clear boundaries, and documentation that stays current.
I choose conventional stacks that can be hired for, draw service boundaries along business lines, and treat runbooks and docs as part of done.
Signals I watch
The measure of a lead is the team that works without them.
I delegate real ownership with context, review for judgement as well as syntax, and create room for engineers to lead a decision end to end.
Signals I watch
Engineering is one voice at the table. I translate between the tradeoffs and the business.
Product gets honest options with costs attached. QA is in planning, not only at the end. Executives get risk in plain language, early.
Signals I watch
I lead a Bay Area co-ed amateur cycling team: racing, recruitment rides, and volunteering at local events. Same job, different peloton. Set the pace, bring new riders in, and make sure nobody gets dropped.