Director's notes

Notes on engineering leadership.

The rest of this site is about the engineering; these essays are about leading it. Each one starts where the learning actually did — with a mistake I made — then works through what it taught me, what I changed, and what happened. They lean on concepts from Will Larson's An Elegant Puzzle and on leading engineering teams, with BlueRobin as the small-scale lab where the same disciplines show up first.

The four states Read

Diagnosing a Stuck Team with the Four States

The quarter I read flat velocity as a headcount problem — and learned it was a clarity problem, not a capacity one.

Running a reorg Read

Running a Reorg Around Domain Boundaries

Organising teams for convenience, then watching cross-zone decisions take days. The reorg around domain boundaries — and Conway's Law — that fixed it.

Sizing teams Read

Sizing Teams in the Copilot Era

Assuming AI assistants meant smaller teams and linear output — and discovering the bottleneck just moved to review and context.

Career ladders Planned

Designing a Career Ladder for Senior ICs

Losing strong engineers because the only ladder ran through management — and building an IC track, with explicit regional expectations, that held.

Org design Read

Organising Teams Around Backlogs, Not Projects

Running a team against three competing backlogs and calling the thrash 'balancing priorities' — until every team got exactly one queue, and priorities fought there instead of for the people.

Culture & values Read

Building a Low-Drama Engineering Team

Tolerating a high performer's toxic behaviour because the output was good. The math was negative — culture is what you tolerate.

Autonomy & trust Read

No News Is Good News: Leading With Trust, Not Status

Chasing status updates to feel in control — and learning that trust-by-default, with the right signals, ships more.

On-call & operational excellence Read

Building a Blameless On-Call

The postmortem that drifted toward 'who broke it', what it cost, and rebuilding on-call into something engineers want to join.

Escalation & candor Read

Learning to Surface Issues Early

A slip the team had known about for two months reached me five days before launch — because I'd taught them, with one irritated reaction, that raising issues early didn't pay.

Systems thinking Read

Reducing WIP: Why Finishing Beats Starting

A team that looked busy and shipped little. Limiting work in progress — not adding people — fixed throughput.

Systems thinking Read

Using System Design to See the Big Picture

Adding two engineers to the team everyone blamed, and watching nothing improve — the constraint lived two stages downstream, and only drawing the system found it.

Metrics & baselines Planned

Managing to Cycle Time, Not Velocity

Chasing a story-point target until quality slipped. Measuring cycle time and stability instead, and letting velocity follow.

Protecting Your Team by Learning to Say No

Saying yes to everything to be helpful — and passing the thrash straight to the team. Becoming a filter, not a funnel.

Shared context Read

Replacing Status Meetings with Observability

The standing meeting that ate hours and decided nothing — replaced with team-health dashboards and an async protocol.

Stakeholder partnership Read

Building a Real Engineering–Product Partnership

Letting engineering settle into an order-taker relationship with product — and rebuilding it into shared ownership of outcomes.

Managing technical quality Read

Fitting Engineering Investment Into a Product Roadmap

Letting 'tech debt later' lose every call against features — until it forced an expensive reckoning. Making it a planned, sized line.

Program sequencing Read

Aligning Cross-Team Features With Incremental Milestones

Coordinating a cross-team feature around one big-bang date. Dependency-aware, incremental milestones fixed the slips.

Migrations Read

Migrations, Not Rewrites: The Only Change That Scales

The rewrite I nearly committed to, and why strangler-fig migrations became my default for changing a codebase — and an org.

Hiring & onboarding Planned

Removing Yourself as the Onboarding Bottleneck

Being the person everyone routed questions to — and turning runbooks, ADRs and retros into something the team could query instead.

Reference: An Elegant Puzzle — Systems of Engineering Management, Will Larson.

Leadership, grounded in practice

These notes sit alongside the measurable delivery documented across the rest of the site.

Explore the projects