--- schema: 1 type: video title: >- How to Build Quality Gates into Agentic Coding Workflows — Nnenna Ndukwe, Qodo AI description: >- A small refund feature passes through planning, implementation, local checks and pull request review before it is considered complete. Nnenna Ndukwe uses SignalPay, a deliberately narrow FastAPI payments application, to turn that sequence into quality gates for an AI coding agent. Participants choose a refund or capture retry workflow, identify the relevant payment rules and create both an implementation plan and a build execution plan. The point is to give the agent an explicit definition of the expected behavior, constraints, verification commands and evidence it must produce. Ndukwe installs repository skills, connects Qodo and walks through the planning templates in Codex. The templates carry assumptions, risk, skill routing, acceptance criteria and the state of the current branch into implementation. After the agent opens a pull request, Qodo adds review findings informed by the codebase and organizational rules. The workshop shows how to inspect those findings, document fixes or deferrals and use a resolver skill to address the feedback and rerun verification. Audience questions test the boundaries: how rules are inferred and approved, how severity controls merge blocking, when verification should run again and how local checks relate to cloud review. The exercise finishes by resolving the remaining issues rather than treating a passing local run as the end of the workflow. Its reusable artifact is a structured process that keeps plans, code changes and review evidence connected. Speaker info: - https://www.linkedin.com/in/nnenna-ndukwe/ - https://github.com/nnennandukwe - https://github.com/nnennandukwe/qodo-quality-workshop-signalpay Timestamps: 0:00 - Structured AI coding workflows 3:15 - Where code quality breaks down 4:09 - Define and codify standards 5:06 - Operationalize context 6:00 - The SignalPay workshop repository 7:11 - Planning, testing, and review gates 8:21 - Repository structure and local setup 10:00 - Quality checkpoints and shared rules 11:32 - Plans and verification evidence 13:23 - Connect Qodo to the repository 16:55 - Choose a payment workflow 18:13 - Prepare Codex with rules and skills 20:47 - Plan the change before implementation 25:07 - Walk through planning templates 27:13 - Build execution and completion notes 28:38 - Implement the refund workflow 29:07 - Skills and local code review 33:17 - Rule categories and severity 36:14 - Inferred rules and human approval 39:10 - Open the pull request for review 42:24 - Verification and review context 44:00 - Rule violations and merge blocking 45:46 - Determinism and rule selection limits 47:13 - Inspect the review findings 49:03 - Resolve issues with an agent skill 50:16 - Rerun verification and update the PR 53:08 - Reuse the workshop process 55:15 - Questions on context and deployment 56:39 - Verify the completed fixes --- Watch along with the corrected transcript, summary, timestamps and resources on this talk's official AIE page: https://ai.engineer/talks/-AKPRSYccJc-youtube-9afe3772fd9912aeceb8 Join our upcoming AI Engineer conferences (see https://ai.engineer landing page): - AIEi Shanghai, November 5–6 — the first for meeting top Chinese labs and engineers - AIE CODE in SF, November 10–12 — the top AI Coding conference returns - AIEi Sydney, December 7–8 — the first AIEi run alongside NeurIPS 2026 in Sydney --- canonical: 'https://liquidgold.cc/watch/-AKPRSYccJc' markdown: 'https://liquidgold.cc/watch/-AKPRSYccJc.md' llms: 'https://liquidgold.cc/watch/-AKPRSYccJc/llms.txt' parent: 'https://liquidgold.cc/channel/aidotengineer.md' channel: aidotengineer updated: '2026-10-23T22:19:14.459Z' language: null children: [] videoId: '-AKPRSYccJc' durationSeconds: 3465 publishedAt: '2026-10-11T16:00:29.000Z' searchChunks: - ordinal: 0 text: null startSeconds: 197 endSeconds: 208 watch: 'https://liquidgold.cc/watch/-AKPRSYccJc?t=197' heldReason: rights - ordinal: 1 text: null startSeconds: 252 endSeconds: 263 watch: 'https://liquidgold.cc/watch/-AKPRSYccJc?t=252' heldReason: rights - ordinal: 2 text: null startSeconds: 263 endSeconds: 306 watch: 'https://liquidgold.cc/watch/-AKPRSYccJc?t=263' heldReason: rights - ordinal: 3 text: null startSeconds: 316 endSeconds: 364 watch: 'https://liquidgold.cc/watch/-AKPRSYccJc?t=316' heldReason: rights - ordinal: 4 text: null startSeconds: 437 endSeconds: 449 watch: 'https://liquidgold.cc/watch/-AKPRSYccJc?t=437' heldReason: rights - ordinal: 5 text: null startSeconds: 449 endSeconds: 507 watch: 'https://liquidgold.cc/watch/-AKPRSYccJc?t=449' heldReason: rights - ordinal: 6 text: null startSeconds: 618 endSeconds: 698 watch: 'https://liquidgold.cc/watch/-AKPRSYccJc?t=618' heldReason: rights - ordinal: 7 text: null startSeconds: 710 endSeconds: 813 watch: 'https://liquidgold.cc/watch/-AKPRSYccJc?t=710' heldReason: rights - ordinal: 8 text: null startSeconds: 1255 endSeconds: 1266 watch: 'https://liquidgold.cc/watch/-AKPRSYccJc?t=1255' heldReason: rights - ordinal: 9 text: null startSeconds: 1523 endSeconds: 1633 watch: 'https://liquidgold.cc/watch/-AKPRSYccJc?t=1523' heldReason: rights - ordinal: 10 text: null startSeconds: 1762 endSeconds: 1997 watch: 'https://liquidgold.cc/watch/-AKPRSYccJc?t=1762' heldReason: rights - ordinal: 11 text: null startSeconds: 1997 endSeconds: 2009 watch: 'https://liquidgold.cc/watch/-AKPRSYccJc?t=1997' heldReason: rights - ordinal: 12 text: null startSeconds: 2187 endSeconds: 2358 watch: 'https://liquidgold.cc/watch/-AKPRSYccJc?t=2187' heldReason: rights - ordinal: 13 text: null startSeconds: 2647 endSeconds: 2658 watch: 'https://liquidgold.cc/watch/-AKPRSYccJc?t=2647' heldReason: rights - ordinal: 14 text: null startSeconds: 2844 endSeconds: 2948 watch: 'https://liquidgold.cc/watch/-AKPRSYccJc?t=2844' heldReason: rights - ordinal: 15 text: null startSeconds: 3030 endSeconds: 3196 watch: 'https://liquidgold.cc/watch/-AKPRSYccJc?t=3030' heldReason: rights - ordinal: 16 text: null startSeconds: 3207 endSeconds: 3322 watch: 'https://liquidgold.cc/watch/-AKPRSYccJc?t=3207' heldReason: rights --- # How to Build Quality Gates into Agentic Coding Workflows — Nnenna Ndukwe, Qodo AI A small refund feature passes through planning, implementation, local checks and pull request review before it is considered complete. Nnenna Ndukwe uses SignalPay, a deliberately narrow FastAPI payments application, to turn that sequence into quality gates for an AI coding agent. Participants choose a refund or capture retry workflow, identify the relevant payment rules and create both an implementation plan and a build execution plan. The point is to give the agent an explicit definition of the expected behavior, constraints, verification commands and evidence it must produce. Ndukwe installs repository skills, connects Qodo and walks through the planning templates in Codex. The templates carry assumptions, risk, skill routing, acceptance criteria and the state of the current branch into implementation. After the agent opens a pull request, Qodo adds review findings informed by the codebase and organizational rules. The workshop shows how to inspect those findings, document fixes or deferrals and use a resolver skill to address the feedback and rerun verification. Audience questions test the boundaries: how rules are inferred and approved, how severity controls merge blocking, when verification should run again and how local checks relate to cloud review. The exercise finishes by resolving the remaining issues rather than treating a passing local run as the end of the workflow. Its reusable artifact is a structured process that keeps plans, code changes and review evidence connected. Speaker info: - https://www.linkedin.com/in/nnenna-ndukwe/ - https://github.com/nnennandukwe - https://github.com/nnennandukwe/qodo-quality-workshop-signalpay Timestamps: 0:00 - Structured AI coding workflows 3:15 - Where code quality breaks down 4:09 - Define and codify standards 5:06 - Operationalize context 6:00 - The SignalPay workshop repository 7:11 - Planning, testing, and review gates 8:21 - Repository structure and local setup 10:00 - Quality checkpoints and shared rules 11:32 - Plans and verification evidence 13:23 - Connect Qodo to the repository 16:55 - Choose a payment workflow 18:13 - Prepare Codex with rules and skills 20:47 - Plan the change before implementation 25:07 - Walk through planning templates 27:13 - Build execution and completion notes 28:38 - Implement the refund workflow 29:07 - Skills and local code review 33:17 - Rule categories and severity 36:14 - Inferred rules and human approval 39:10 - Open the pull request for review 42:24 - Verification and review context 44:00 - Rule violations and merge blocking 45:46 - Determinism and rule selection limits 47:13 - Inspect the review findings 49:03 - Resolve issues with an agent skill 50:16 - Rerun verification and update the PR 53:08 - Reuse the workshop process 55:15 - Questions on context and deployment 56:39 - Verify the completed fixes --- Watch along with the corrected transcript, summary, timestamps and resources on this talk's official AIE page: https://ai.engineer/talks/-AKPRSYccJc-youtube-9afe3772fd9912aeceb8 Join our upcoming AI Engineer conferences (see https://ai.engineer landing page): - AIEi Shanghai, November 5–6 — the first for meeting top Chinese labs and engineers - AIE CODE in SF, November 10–12 — the top AI Coding conference returns - AIEi Sydney, December 7–8 — the first AIEi run alongside NeurIPS 2026 in Sydney --- Canonical: https://liquidgold.cc/watch/-AKPRSYccJc · Channel: [AI Engineer](https://liquidgold.cc/channel/aidotengineer.md) > Search chunk text is held until public-source provenance is verified. ## Document chunks > Missing: no stored document chunks are available.