Developer
Mission
Section titled “Mission”Build working, tested software that meets the acceptance criteria and the team’s quality standards. Developers turn designed tasks into a merged, deployable increment.
Responsibilities
Section titled “Responsibilities”- Implement features against acceptance criteria and the technical design.
- Write unit and integration tests alongside the code.
- Keep code within the team’s standards (style, security, patterns).
- Open small, reviewable pull requests and respond to review feedback.
- Fix defects found in review and testing.
- Update code-level documentation and contribute to ADRs when they make decisions.
Inputs
Section titled “Inputs”| Input | From |
|---|---|
| Designed, estimated tasks | Architect/Dev Lead |
| Stories & acceptance criteria | BA / Stakeholders (Product Owner) |
| Coding & quality standards | Architect/Dev Lead, Quality Engineer |
| Review feedback | Quality Engineer |
Outputs
Section titled “Outputs”| Output | To |
|---|---|
| Implemented features + unit tests | Quality Engineer (via PR) |
| Pull requests | Quality Engineer |
| Updated technical docs | Team |
| Defect fixes | Quality Engineer |
AI tools & how we use them
Section titled “AI tools & how we use them”This is the most AI-intensive role. AI pair-programming is the default way of working, with the developer as reviewer and decision-maker. For how this scales into autonomous agents and spec-driven workflows, see Agentic & Spec-Driven Development.
- GitHub Copilot / Cursor / Claude Code (in-editor pair programming): autocomplete, multi-file edits, “explain this code,” and generating implementations from a described task. Strong fit for .NET/C# and IaC work.
- Claude / ChatGPT (problem-solving & refactoring): rubber-duck a tricky bug, propose a refactor, translate between frameworks, or generate boilerplate. Prompt: “Refactor this C# service to use the repository pattern; keep behavior identical and add unit tests.”
- AI unit-test generation: generate test cases including edge cases from the implementation and acceptance criteria; the developer prunes and asserts correctness.
- AI code explanation / onboarding: understand unfamiliar parts of the codebase quickly (“explain what this module does and its dependencies”).
- Inline security hints (Copilot, Snyk in-IDE): catch obvious vulnerabilities while typing.
flowchart LR
T[Task + ACs] --> AI[AI drafts code + tests]
AI --> D[Developer reviews & corrects]
D --> LOCAL{Tests pass locally?}
LOCAL -->|no| D
LOCAL -->|yes| PR[Open small PR]
PR --> SA[Solutions Architect review]
Collaboration
Section titled “Collaboration”Developers take design and tasks from the Project Manager, clarify requirements with the BA, submit PRs to the Solutions Architect/Dev Lead, and fix issues they raise. They self-organize on how to deliver the sprint backlog.
Definition of Done for this role
Section titled “Definition of Done for this role”- Code meets acceptance criteria and team standards.
- Unit/integration tests written and passing; coverage threshold met.
- PR is small, reviewed, and approved by the Quality Engineer with no blocking comments.
- CI (build, tests, security/quality gates) is green; docs updated.