Skip to content

Add Learning Path: AI coding agents on Google Axion - #4023

Open
ranimandepudi wants to merge 1 commit into
ArmDeveloperEcosystem:mainfrom
ranimandepudi:codex/axion-agent-studio-learning-path-review
Open

ranimandepudi wants to merge 1 commit into
ArmDeveloperEcosystem:mainfrom
ranimandepudi:codex/axion-agent-studio-learning-path-review

Conversation

@ranimandepudi

@ranimandepudi ranimandepudi commented Nov 6, 2026 •

Copy link
Copy Markdown
Contributor

This Learning Path shows how hosted Gemini planning and three local Gemma coding agents on Google Axion turn a bounded checkout request into a tested application change. It explains GKE Agent Sandbox isolation, C4A/N4A workload placement, independent validation, and the difference between VM capacity, Pod limits, and measured CPU usage.

This PR is ready for editorial and technical team review. It is not ready to merge for publication: the source-release and clean-cloud walkthrough requirements below remain open. The index and child pages retain draft: true.

Scope

  • Nine Markdown pages covering prerequisites, configuration, deployment, running and verifying a change, CPU interpretation, and cleanup.
  • Editable technical architecture SVG and high-resolution PNG.
  • Documentation and diagrams only. The application source archive is excluded because its public release and license are still pending. The preparation page identifies this dependency explicitly, without a broken download link.

Validation

  • Repository metadata checker and staged whitespace check pass.
  • All nine pages parse; metadata, navigation weights, descriptions, draft cascade, and code fences checked.
  • Twenty Bash blocks and two embedded Python programs pass syntax checks.
  • The exact documentation-only PR content passes a draft-inclusive Hugo extended 0.130.0 build using the upstream Arm theme and a local preview overlay.
  • GitHub CI passed for commit ceeb005: the repository Hugo build, content path checks and metadata checks succeed. The normal Hugo build excludes draft pages; the separate local build includes them.
  • The snippet test suite reports maintenance off and skips runtime execution. This CI success is not evidence of a live GKE walkthrough.
  • The earlier local reading preview passed page/asset loading and desktop/narrow visual checks. Independent editorial, technical-source, and publication-rule reviews informed the corrections.

Required before publication

  • Approve the companion application's public release, license, and notices; add an immutable source location and recheck the instructions against it.
  • Complete a fresh Linux Arm64/GKE walkthrough from building images through real Gemini planning, local Gemma inference, validation, interactive checkout checks, metrics, and scoped cleanup.
  • Complete Learning Path team review and confirm applicable contribution licensing before merge/publication.

The original demo has recorded Arm runtime evidence, but that recording and local fixture tests do not establish clean-environment reproducibility of the portable instructions. No new cloud deployment was performed for this PR.

Please review the learning outcome, prerequisite scope, instruction clarity, architecture accuracy, and Arm resource explanation. The remaining source-release and runtime checks are intentionally open.

Contributor review

  • Reviewed Create a Learning Path.
  • Checked the submitted documentation and images for confidential information. Private deployment configuration, account details, recorded run data, and credentials are excluded.

The PR template's contribution-license confirmation remains for contributor/maintainer review before merge. This documentation review does not supply a license for the excluded application source.

@ranimandepudi
ranimandepudi marked this pull request as ready for review November 6, 2026 15:47
@ranimandepudi

Copy link
Copy Markdown
Contributor Author

Hi @jasonrandrews and @anupras-mohapatra-arm, could you please review this Learning Path?

Jason, your feedback on the Axion/GKE architecture and technical workflow would be appreciated. Anupras, could you review the instruction clarity, structure, and alignment with Learning Path writing standards?

The PR is ready for team review and the static checks pass. The companion-source release approval and clean-cloud walkthrough remain open before publication, as documented in the PR description.

My account cannot add formal reviewer assignments to this repository. Could a maintainer add you both to the reviewers list? Thank you.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant