One-Year AI Curriculum
A structured 12-month plan aligned with Coursera Plus, IBM AI programs, and the goal of building a real, local-first AI orchestration system.
High-level phases
Phase 1 · Foundations (Months 1–2)
Refresh Python and math, establish fast study habits, and lay the groundwork for the more advanced IBM / DeepLearning.AI tracks.
- Python for Everybody Specialization
- Mathematics for Machine Learning
- IBM intro-level generative AI overview (for terminology only)
Phase 2 · Core ML Engineering (Months 3–6)
Build the machine learning backbone: classic models, evaluation, and practical engineering patterns.
- IBM Machine Learning Professional Certificate
- IBM AI Engineering Professional Certificate
- Optional: begin watching Karpathy’s Neural Networks: Zero to Hero series to build intuition for the later LLM-from-scratch work.
Phase 3 · Transformers, LLMs & Systems (Months 7–9)
Move into generative AI, transformers, LLMs, and system-level design, anchored by a from-scratch implementation path.
- Andrej Karpathy – “Neural Networks: Zero to Hero” + GPT-from-scratch implementation (core hands-on LLM track).
- IBM Generative AI Professional Certificate
- DeepLearning.AI courses on LLMs and “Building systems with LLMs”
Phase 4 · Deployment & On-Device (Months 9–11)
Focus on shipping: Core ML, on-device inference, Docker, and cloud/on-prem orchestration.
- Machine Learning on Devices / on-device AI courses
- Cloud/DevOps for ML fundamentals (Docker, containers, pipelines)
Phase 5 · Capstone (Month 12)
Build and document the first production-grade version of the MEWX.AI orchestration system.
- Design, implement, and benchmark the orchestration stack
- Write public-facing docs and internal lab notes