docs: add MLOps-with-coding-agents how-to

[typhoonzero]

Summary

Follow-on to #245. Adds a second how-to: Run MLOps with Coding Agents and On-Premise LLMs under model_inference/inference_service/how_to. Stacked on the PR1 branch — once #245 lands, GitHub will auto-retarget this PR to master.

The doc focuses on four agent-driven MLOps workflows on Alauda AI:

• Manage InferenceService / LLMInferenceService — agent loop (draft → server-side dry-run → apply → poll → smoke-test), with concrete starter prompts.
• Configure Envoy AI Gateway — auth and rate limits via AIGatewayRoute, AIServiceBackend, BackendSecurityPolicy, SecurityPolicy, and BackendTrafficPolicy; links to the existing intro/install docs.
• Tune service performance to fit hardware — explicit five-step agent-driven loop (SLOs → benchmark → one-change-at-a-time → measure → stop on SLO or ceiling); cross-links to the prior doc's tuning section instead of duplicating it.
• Plan fine-tuning and generate reports — a tool-selection table (Notebook / Training Hub / Kubeflow Trainer v2 / LLM Compressor) and two ready-to-commit markdown templates: a pre-run plan and a post-run report, both designed to be filled in by the agent from live cluster data.

Also includes a "daily MLOps loop" walk-through, guardrails (read-only first, --dry-run=server, one change per iteration, never fabricate metrics, no hosted-provider fallback), and a References section.

The amended PR1 commit adds a \{#best-practices} anchor to the prior doc so this doc's deep links resolve under doom lint.

Test plan

• doom lint passes on the new file (0 errors, 0 warnings)
• Pre-commit yarn lint passes
• All internal cross-links (./*.mdx, ../../../envoy_ai_gateway/*, ../../../kubeflow/how_to/*, ../../../workbench/how_to/*, ../../../llm-compressor/how_to/*, ../../../infrastructure_management/hardware_profile/*) resolve to files that exist on the stacked base
• Reviewer: verify rendered page and the two markdown templates render cleanly inside fenced code blocks

🤖 Generated with Claude Code

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

🗂️ Base branches to auto review (3)

• main
• master
• ^\d.x$

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 9bb43dd8-8527-4a9c-b3d2-73bdf6dfcc38

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch docs/mlops-with-coding-agents

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[typhoonzero]

Superseded — both docs are now in #245. Closing.