Module 10: Start Here

Overview

Everything so far produced a design, a practice, and the standing to get them agreed. This module is about the condition you will actually work in, and it closes the guide.

How does architecture change over time, inside constraints you do not control — and what is the job of changing it becoming?

Greenfield architecture is a training exercise. The real work is changing a system that is running, earning revenue, and staffed by people with opinions, without the capacity you would need to do it properly. That is not an unusual situation; it is the normal one.

The organising idea

One constraint governs the first half:

Every evolution decision is a sequencing decision under fixed capacity.

Refactor or rewrite, repay debt or ship features, buy or build — none is really a question about the best end state. Each is a question about what to do first with engineering time that is already committed. A plan assuming capacity nobody has is a wish with a diagram.

The second half turns the same lens on you. The scarce skill is moving, and the guide’s own record is the evidence:

The scarce skill is no longer producing a design. It is knowing whether a design is wrong.

How the module is built

graph TD
  A["<b>1 · Scale, context<br/>and debt</b><br/>transitions, engagement types,<br/>the quadrant, the payoff line"] --> B["<b>2 · Changing a<br/>running system</b><br/>strangle, sequencing,<br/>legacy, buy vs build"]
  B --> C["<b>3 · What the job<br/>is becoming</b><br/>displacement, liability,<br/>competencies, tracks"]
  C --> D["<b>4 · Your trajectory</b><br/>roles, AI teammates,<br/>the environment, learning"]
  D --> E(["Workbook — Exercise 10"])

  classDef sys fill:#5B8FB9,stroke:#4A7A9E,stroke-width:2px,color:#ffffff
  classDef you fill:#E8A33D,stroke:#C4832A,stroke-width:2px,color:#ffffff
  classDef cap fill:#D9695B,stroke:#B4483A,stroke-width:2px,color:#ffffff
  class A,B sys
  class C,D you
  class E cap

The first half changes the system. The second changes what you do next — and the Workbook’s capstone assembles everything behind you into one package.

Section 1 — Scale, context and debt

1. Evolution with scaleWhat architecture means at 50 engineers, at 150 and at 500.

2. What breaks at the next transitionReading a running migration's rate, and what it makes true.

3. Engagement typesProduct, outsourcing, integration, transformation — and what transfers.

4. Technical debt, preciselyThe quadrant, and the response each kind demands.

5. The design payoff lineWhere borrowing stops accelerating you — and what is not debt at all.

Section 2 — Changing a running system

6. Refactor, rewrite, or strangleThree moves, and what each one requires to work.

7. Sequencing under zero spare capacityShipping value when there is no capacity to borrow.

8. AI applied to legacyWhere it multiplies scarce context — and the trap that defeats it.

9. Buy vs build under a moving frontierWhat is worth owning when the capability changes every quarter.

10. Hard situations and the human texturePlatform against product, inherited decisions, and sustaining the role.

Section 3 — What the job is becoming

11. Threat and opportunityWhat gets displaced, and what gets scarcer.

12. Code as liability, followed throughModule 1's thesis taken to its uncomfortable conclusions.

13. The shifting competency profileWhat is rising — tested against this guide's own findings.

14. Tracks, and the skill ceilingWhy learning more technologies is how people plateau.

Section 4 — Your trajectory

15. Emerging roles, and reading a titleWhat is genuinely new, what is rebranded, and five questions that tell you.

16. Working with AI teammatesWhat delegates, what does not, and where the boundary sits.

17. Configuring the environmentShaping what the whole team produces, rather than reviewing each result.

18. Your learning strategyWhat to learn now, defer, and deliberately skip.

How to work this module

About five hours of reading, plus about four for the optional capstone exercise, which is assembly rather than new design.

1. Read the eighteen topics — about 5 hours.

2. Optional — Workbook exercise 10 — about 4 hours: a debt assessment, a legacy strategy, a buy-vs-build decision, a migration sequence, a competency gap analysis, a learning roadmap, and the final portfolio.

Before you start

Gather every artifact from Modules 1–9. The second half of the capstone exercise is assembly, not production, and the work is in making nine modules of exercises read as one engagement.

Expect contradictions. Your Module 1 stakeholder map was naive by Module 8’s standards. Your Module 2 NFR budget set a trigger that Module 6’s arithmetic invalidated. Your Module 3 strategy treated a dependency as interim that this module shows lasts years. Reconciling or acknowledging those is a rubric item, and it is what makes the portfolio credible rather than tidy.

Check your understanding

Answer now, revisit at the end.

1. Name something in a system you know that gets called technical debt. Was it a deliberate trade, or just mess?
2. A migration you have seen has been “nearly done” for a long time. What was the actual remaining work?
3. Which part of your own work would you be least comfortable having a model do unsupervised?
4. If someone asked for evidence that you can do architectural work, what would you show them today?

→ Start with Evolution with scale.