Why these notes exist

The goal for the next 12–18 months is to own outcomes, not ceremonies: to become the Engineering Manager for a Machine Vision / AI platform, with enough architectural credibility to make and defend real technical decisions.

These notes follow that roadmap one month at a time. Each chapter explains a topic, connects it to an image → annotation → training → deployment pipeline, and ends with questions to check myself.

Sections

SectionWhat it coversStatus
Software ArchitectureStyles, boundaries, quality attributes, C4, ADRs and trade-offsMonth 1 · Nov 2026
System DesignRequirements, estimation, scaling, caching, queues, data distribution, reliabilityMonth 1 · Nov 2026
Python & FastAPIComing in DecemberPlanned
Containers, CI/CD & MongoDBComing in JanuaryPlanned

How to use a chapter

Every chapter has the same shape, so it’s easy to revise:

  1. In one line: the idea, in one line.
  2. The idea and how it works: the explanation, with a diagram where it helps.
  3. On the platform: what it looks like in a machine-vision pipeline (Blob Storage, Azure Functions, Kubeflow, MongoDB, V7).
  4. Trade-offs and common mistakes.
  5. Check yourself: questions with hidden answers. Click to reveal.
  6. Apply it: one small task for the capstone or for work.

When to tick a topic

Tick an item on the roadmap only when I can explain it without notes and have used it once in the capstone or at work.

The rules for the year

  1. One primary resource at a time.
  2. At least half of each week’s hours spent building, not watching.
  3. Every month ends with a deliverable.
  4. YouTube is for specific blockers, not a curriculum.

Month 1 deliverable

  • A current-state architecture diagram of the real platform (C4 context + container)
  • A skills-gap list
  • A first draft of platform KPIs

See C4 Diagrams and Requirements to get started.