Skip to content

Latest commit

 

History

10 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 

Repository files navigation

AI/ML × Cloud × DevOps × MLOps — 12-Stage Practitioner Roadmap (12–24 Months)

A practitioner-verified, text-first engineering curriculum (12–24 Months) grounded entirely in official vendor documentation, university courses, foundational textbooks, and working code. Zero video tutorial fluff.

🌐 Live Interactive Checklist & Curriculum: https://wyrcan-io.github.io/roadmap/


🎯 Why This Roadmap Exists

Most online roadmaps fall into one of two traps:

  1. Tutorial Hell: A list of 40-hour video playlists where you copy code without building real mental models.
  2. Disconnected Toy Repos: Building 20 separate "Hello World" projects that don't reflect production engineering.

This roadmap is built on 4 non-negotiable principles:

  1. 70% Building: Use books and official docs to unblock real implementations. Building is the actual curriculum.
  2. One Evolving System: Build and evolve a single production-grade AI platform across your 12 milestones.
  3. Break Things: Deliberately induce failures, trace errors through telemetry, fix them, and write postmortems.
  4. Design First: Reason methodically about traffic, consistency, latency, cost, and failure modes before writing code.

⏱️ Dual-Track Pacing Guide (12 vs 24 Months)

Mastering 4 canonical textbooks (ISLP, UDL, DDIA, OSTEP), 7 major platform stacks, and 110 checkpoints is cognitively intense. Rather than pretending this is a quick 6-month bootcamp, we provide two explicit, realistic tracks:

Dimension 🛠️ Working Practitioner Track (Recommended) ⚡ Accelerated Sprint Track
Total Duration 24 Months (2 Years) 12 Months (1 Year)
Weekly Commitment 8–10 hours / week (1 hr/day + weekend build block) 20–25+ hours / week (full-time daily commitment)
Pace per Milestone ~8 weeks / 2 months per stage ~4 weeks / 1 month per stage
Target Audience Working engineers, full-time professionals, students Dedicated career sabbaticals / full-time bootcampers
Pedagogical Balance 4 weeks reading & labs + 4 weeks building, failure testing & runbooks 2 weeks reading/labs + 2 weeks architecture provisioning & exit test
System Design (DDIA) 1 chapter every 6–8 weeks (deep retention & synthesis) 1 chapter every 3–4 weeks (high cognitive load)

🗺️ The 12-Stage Curriculum Overview

Stage Track Core Focus Primary Deliverable Recommended Pace
00 Foundation Reset Python Model, Asyncio, OSTEP OS, MDN HTTP, PostgreSQL Trace web request from DNS to DB query plan 6–8 Weeks (Fast: 3–4w)
01 Platform Linux Internals, CLI Fluency, Networking, systemd Deploy hardened FastAPI service on Linux VM behind Nginx 8 Weeks (Fast: 4w)
02 Platform Docker Deep Dive, Multi-Stage Builds, OCI, Compose Production multi-tier compose stack with networking isolation 8 Weeks (Fast: 4w)
03 Platform AWS Architecture, VPC Subnets, IAM, RDS, S3, ALB Provision AWS multi-tier architecture + System Design Doc 8 Weeks (Fast: 4w)
04 Platform GitHub Actions, CI/CD, 12-Factor App, OWASP Security Automated CI/CD pipeline with Trivy scans & zero-downtime rollback 8 Weeks (Fast: 4w)
05 Platform Terraform IaC, Remote State S3/DynamoDB, Reusable Modules Spin up and tear down staging cloud environments via Terraform 8 Weeks (Fast: 4w)
06 AI / ML ML Fundamentals, ISLP Python Labs, scikit-learn MOOC Classical tabular ML pipeline service exposed via FastAPI 8 Weeks (Fast: 4w)
07 AI / ML PyTorch, Autograd, Custom Training Loops, UDL Book Train PyTorch model from scratch + GPU inference API 8 Weeks (Fast: 4w)
08 MLOps Rules of ML, MLflow Tracking, Model Registry, Lineage Reproducible ML training pipeline with versioned artifacts 8 Weeks (Fast: 4w)
09 MLOps Continuous Training (CT), Automated Quality Gates Automated model promotion with canary deployment routing 8 Weeks (Fast: 4w)
10 Platform Kubernetes (EKS / K3s), Pods, Ingress, Helm Charts Package and orchestrate application in Helm across Kubernetes 8 Weeks (Fast: 4w)
11 Production OpenTelemetry, Prometheus, Grafana, Drift & SLOs Telemetry dashboards + induced failure incident postmortem 8 Weeks (Fast: 4w)
12 Capstone Architecture Hardening, C4 Diagrams, Well-Architected Review Production portfolio platform package & security audit 8 Weeks (Fast: 4w)
System Design DDIA 2nd Edition, Microsoft Cloud Patterns, Builders' Library Parallel continuous architecture loop (Stages 01–12) Continuous

⚡ The Non-Negotiable 13-Pillar Backbone

If the entire catalog feels vast, these 13 items form your core intellectual spine:

  1. Foundation: MIT Missing Semester + OSTEP + Pro Git + MDN HTTP
  2. Linux & Networking: The Linux Command Line + Beej's Guide to Network Concepts
  3. Docker: University of Helsinki DevOps with Docker + Docker Build Best Practices
  4. AWS: AWS Skill Builder + Well-Architected Framework + Amazon Builders' Library
  5. DevOps: GitHub Actions + Google SRE Books
  6. Terraform: HashiCorp AWS Tutorials + Modular IaC
  7. ML Fundamentals: An Introduction to Statistical Learning (ISLP) + Google MLCC + scikit-learn MOOC
  8. Deep Learning: PyTorch Basics + Understanding Deep Learning (Prince / MIT Press)
  9. ML Engineering: Google's Rules of ML + Hidden Technical Debt in ML (NeurIPS) + MLflow
  10. MLOps: Google Cloud MLOps Guide + DataTalksClub MLOps Zoomcamp
  11. Kubernetes: Kubernetes Official Tutorials + Helsinki K8s MOOC + Amazon EKS Workshop
  12. Observability: OpenTelemetry Primer + Prometheus + Evidently AI
  13. System Design (Continuous): Designing Data-Intensive Applications (DDIA 2nd Edition) + Microsoft Cloud Design Patterns + Amazon Builders' Library + C4 Model

🛠️ Interactive Features on the Live Site

The live interactive site includes:

  • Google Docs-style live Outline panel with scrollspy tracking.
  • 110 Granular Checkpoints saved locally in browser storage.
  • Warm Dark Mode (Espresso) & Light Mode (Warm Ivory) toggle.
  • Export Progress to Markdown for Notion / Obsidian / personal dev journals.
  • Live Search Filter across all 59 primary sources and documentation chapters.

🏛️ Maintained By

Maintained with care by Wyrcan as a 100% free and open public resource for engineers worldwide.

About

A practitioner-verified, text-first 12-month engineering roadmap for AI/ML, Cloud, DevOps, and MLOps grounded in official docs and canonical books.

Topics

Resources

Stars

1 star

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages