Top 100 AI Platform interview questions with Shikha Singh

Top 100 Real-Life Interview Questions — AI Platform & AI Infrastructure

Every question in this ebook was asked in a real interview room: 15 rounds, from screening calls to a VP round. None are invented to pad the list.

It's written for engineers moving from DevOps or Cloud into AI Platform and AI Infrastructure roles. It covers the platform fundamentals interviewers still test, plus the GPU and LLM-serving questions they've added.

What you get

• 100 real interview questions across 10 topics
• 100 hands-on drills, one short task per question, so you can survive the follow-up
• 84-page PDF
• Senior-level answers that explain the mechanism, not just the tool name, with code and config where it helps

Topics covered

Containers, Linux & CI/CD · GitOps, Helm & Progressive Delivery · Terraform · Kubernetes Core & Operations · AWS Networking, Storage & EKS · Cloud Architecture & Edge · Security & IAM · Databases · Observability · GPU & LLM Infrastructure (MIG, time-slicing, vLLM, quantization, KEDA, NVIDIA drivers)

Sample questions

• A 24 GB GPU and a 2.4 GB model: how many models fit?
• How do you quantize a 30B model, and what's the memory math?
• Pods fail with "IP not available" while the nodes look healthy. Fix it.
• Your GitOps rollback target isn't in the registry or in Git. Recover it.

Why this ebook

Most candidates don't fail on knowledge. They name the right tool and miss the mechanism, or they go silent on a tool they haven't used. Every answer here is written to survive the follow-up question.

About the author

Shikha Singh is a Senior AI Platform Engineer with 11+ years of experience and holds the CKA and CKS certifications. She built this ebook from her own interview debriefs, breaking each concept on purpose until it stuck.

Instant download · PDF · For personal use only