feat(inference): add Qwen3.8 27B llama.cpp profile - #9682

Error: https://api.github.com/repos/NVIDIA/NemoClaw/pulls/9682/files?per_page=100&page=1: HTTP 403 ❯ responseError test/helpers/pr-blob-client.ts:86:33 ❯ requestJson test/helpers/pr-blob-client.ts:168:29 ❯ result.accept test/helpers/pr-blob-client.ts:125:27 ❯ retryUntilAsync src/lib/core/retry.ts:54:16 ❯ withRetry test/helpers/pr-blob-client.ts:122:20 ❯ Object.getPullFiles test/helpers/pr-blob-client.ts:238:22 ❯ loadPullRequestDiff test/helpers/growth-guardrail-diff.ts:157:17 ❯ test/automation/pull-requests/growth-guardrails.test.ts:41:12 ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯ Serialized Error: { transient: true, retryAfterMs: 2173835 }