devops devopssreplatform-engineeringroadmapskuberneteschaos-engineeringai-infra

devops-sre-portal-architect

Designs and generates 2026-2027 DevOps, SRE, Platform Engineering, and AI/GPU infrastructure roadmaps, production Kubernetes/bare-metal manifests, chaos troubleshooting labs, and AI-Coder learning-to-career portal modules. Use for curriculum design, platform blueprints, and SRE competency matrices.

Quality 100% Code blocks 5 Sections 8 Lines 102 RU translation yes
view source
Models: sonnetopusgpt-6gemini-3glm-5

DevOps & SRE Portal Architect

Progressive competency tracks from Linux internals to bare-metal Talos, GPU inference, and chaos drills.

Quick Start

# Validate all roadmap tracks and skill definitions in the repository
python scripts/validate_roadmaps.py
python -m claude_skills.cli validate --dir .claude/skills
# Check roadmap track coverage and live stage counts
import json
from pathlib import Path

tracks = [json.loads(p.read_text(encoding="utf-8")) for p in Path("roadmaps").glob("*.json")]
live = [t["slug"] for t in tracks if t.get("status") == "live"]
print(f"Live tracks ({len(live)}/{len(tracks)}): {live}")

When to Use

  • Structuring Junior-to-Principal DevOps, SRE, Platform Engineering, and AI/LLMOps roadmaps (Maps 01-23)
  • Pairing infrastructure theory with runnable Kubernetes manifests, Go operators, and Chaos Drill troubleshooting labs
  • Building or updating the AI-Coder Learning, Practice, and Career Portal on top of ssrjkk/agent-skills
  • Designing System Design architectures (10k+ RPS, Multi-AZ) and ATS-ready XYZ resume achievements

Best Practices

  • Organize learning stages progressively: Foundation (01-03), Cloud-Native & SRE (04-13), Principal & AI-Infra (14-20), and Vibe-Coding/MCP/SEO (21-23).
  • Never recommend deprecated patterns for new deployments: use Kafka KRaft instead of ZooKeeper, Cilium eBPF instead of sidecar-heavy meshes, and OIDC federation instead of static CI keys.
  • Pair every roadmap stage with a concrete artifact: a hardened YAML manifest, a Go reconciler, a Dagger pipeline, or a chaos verification script.
  • Keep English and Russian roadmap stages 1-to-1 in item count so validate_roadmaps.py passes without drift.
  • Connect technical mastery directly to career outcomes (Learn -> Build -> Get Hired) using verifiable GitHub portfolio blueprints.
  • Enforce SSR/prerendering and Open Graph metadata on all generated portal pages so search engines index every track.

Dependencies

# Core validation and site generation tools
python3 -m pip install -e ".[dev]"
python3 scripts/validate_roadmaps.py

Examples

{
  "$schema": "../roadmaps.schema.json",
  "slug": "devops",
  "name": "DevOps & SRE",
  "name_ru": "DevOps и SRE",
  "icon": "🚀",
  "tagline": "From Linux kernel and eBPF to Talos bare-metal, vLLM GPU clusters, and SRE chaos drills",
  "tagline_ru": "От ядра Linux и eBPF до bare-metal Talos, GPU-кластеров vLLM и SRE траблшутинга",
  "status": "live",
  "eta": "2026-11",
  "catalog_domain": "devops"
}
# Verify 1-to-1 bilingual parity across all stages of a roadmap track
import json
from pathlib import Path

track = json.loads(Path("roadmaps/devops.json").read_text(encoding="utf-8"))
for idx, stage in enumerate(track["stages"]):
    assert len(stage["items"]) == len(stage["items_ru"]), f"Mismatch at stage {idx}"
    assert len(stage["items"]) >= 2
print(f"Verified {len(track['stages'])} bilingual stages successfully.")

Step-by-Step

1. Audit existing tracks in roadmaps/*.json and skills in .claude/skills/ to identify coverage gaps.

2. Map competencies across the 4 progressive tiers: Foundation, Cloud-Native/SRE, Principal/AI-Infra, and Vibe-Coding/Portal.

3. Draft bilingual stage titles (title, title_ru) and 1-to-1 item lists (items, items_ru) with concrete tools and metrics.

4. Link existing catalog skills in related_skills and set catalog_domain to a valid domain from skills_catalog.json.

5. Bundle hands-on labs (break-my-cluster.yaml, verify-fixes.sh, vLLM and CloudNativePG manifests) alongside the tracks.

6. Run python scripts/validate_roadmaps.py and python scripts/build_docs.py to regenerate the bilingual portal.

7. Verify that README.md and docs/roadmaps.html reflect the updated live roadmap counts.

Validation

1. python scripts/validate_roadmaps.py reports 0 errors and confirms all live tracks have full RU mirrors

2. Every skill listed in related_skills exists in skills_catalog.json

3. python scripts/check_site_consistency.py passes across all generated HTML pages

4. All accompanying Kubernetes, Go, and Python lab files pass syntax and schema validation

Troubleshooting

  • items_ru length mismatch error — ensure every English item in a stage has a corresponding Russian item at the exact same index.
  • Unknown related_skills failure — run python -m claude_skills.cli catalog before validating roadmaps so new skills are indexed.
  • Unknown catalog_domain failure — verify the domain has at least one skill in .claude/skills/<domain>/ or set catalog_domain to null.
  • Russian header not toggling on roadmaps.html — confirm both name_ru and tagline_ru are populated on the live track JSON.