AI Infra Interviews logo

The AI infra curriculum

The question bank answers what to practise. These answer what to learn, and in what order. Each course is a reading path through the work a AI infrastructure engineer actually does, with the questions and concepts that go with it. Free lessons in every course, no card needed.

Grounded in real AI infrastructure interview loops and written to a senior-engineer editorial bar, with every number worked and every diagram hand-built.

START HERE, FREEA GPU is a bandwidth machine that happens to do arithmetic12 min · no account needed→

The main path

Work these in order.
beginner5 modules · 15 lessons · ~7 hours5 free to read

Foundations of AI Infrastructure

What a GPU actually is, what happens between a prompt and a token, and how to size a model service before you deploy it. The trunk course every AI infrastructure path starts from, for engineers who can ship software but have never run a GPU in production.

  1. 01The machine3/3 free
  2. 02The request2/3 free
  3. 03The arithmetic3
  4. 04The service3
  5. 05The deployment3
READ FREEA GPU is a bandwidth machine that happens to do arithmetic12 min →
advanced5 modules · 16 lessons · ~8.5 hours2 free to read

Inference Engineering

The mechanisms that decide throughput, time to first token and cost per token in a model serving fleet: how the scheduler and the cache allocator actually work, what to measure, and which lever to pull when a number moves the wrong way.

  1. 01The engine2/3 free
  2. 02The cache3
  3. 03The scheduler3
  4. 04The precision3
  5. 05The fleet4
READ FREEA serving engine is a scheduler with a memory allocator attached13 min →
advanced4 modules · 12 lessons · ~7.5 hours1 free to read

Training at Scale

Planning, running and rescuing training jobs large enough that failure is a schedule item: what a step actually costs, how memory decides the parallel layout, what the collectives are worth, and what to do when a run degrades or diverges.

  1. 01The run1/3 free
  2. 02The layout3
  3. 03The Economics3
  4. 04The Rescue3
READ FREEA training step is four costs and only one of them is the model13 min →
advanced4 modules · 12 lessons · ~7.5 hours1 free to read

Fleet and Reliability

Running the machine that other people's jobs run on: how a GPU gets scheduled and why fragmentation is the real cost, the four networks and what each one breaks, how devices fail and what catches it, and how to share a cluster between tenants without lying to any of them.

  1. 01The scheduler1/3 free
  2. 02The fabric3
  3. 03The fleet3
  4. 04The tenancy3
READ FREENinety percent allocated and thirty percent used is the normal failure14 min →
ALONGSIDE THE COURSES
The must-know questions →The concept curriculum →Interview prep by company →All AI infra interview questions →