07 / APPROACH

How an engagement runs

Six phases. Each one designed to be stopped without leaving a mess. That's intentional.

Discover

1–2 weeks

We read the code, the incident history, and the architecture docs nobody's looked at in two years. We talk to the people who built it and the people who operate it. We find the constraints that aren't in the slide deck.

Outputs: Constraint map Failure mode catalog Current-state assessment

Architect

1–3 weeks

Boundaries, data contracts, threat model, and a delivery sequence that can be stopped at any point without leaving the system in a broken state. We design the exit before we design the feature.

Outputs: Architecture decision records Service boundary map Delivery sequence

Engineer

2–16 weeks

Thin vertical slices. Production-like environments from week one. Interfaces frozen before internals are clever. We don't build the whole thing and then find out the API was wrong.

Outputs: Working software API contracts Deployment pipelines

Validate

1–2 weeks

Load testing, failure injection, security review, and parity tests against the system being replaced. "It works on my machine" is not a gate. We break it on purpose so production doesn't have to.

Outputs: Load test results Security review Parity test suite

Deploy

1–2 weeks

Progressive delivery. Rollback that has been rehearsed, not just documented. On-call that includes the people who wrote the code. We don't throw it over the wall.

Outputs: Production deployment Runbook Rollback procedure

Optimize

2–4 weeks

Latency budgets, cost optimization, and operability improvements after the applause. We stay through the first real traffic, not just the launch. The system isn't done when it deploys — it's done when it runs smoothly.

Outputs: Performance baseline Cost analysis Operability handoff

Ready to start?

Every engagement starts with a 45-minute architecture call. We'll figure out if the work is in our range.

Schedule a briefing