AI/ML Engineer Interview Playbook: 141 Questions with Mohd Ali Naqvi

141 questions with clear, interview-ready answers, tagged Basic, Intermediate, Scenario and Tricky

18 curveball scenarios that test hands-on experience, such as:

1. 1. RAG that works in the demo but fails in production
   2. Different outputs at temperature 0
   3. Agents stuck in loops
   4. Prompt injection hidden in a PDF
   5. A model with 99% accuracy that is useless

Interviewer’s trap” notes showing the common wrong answer and the follow-up to expect

Working code for live-coding and ML-from-scratch questions: logistic regression, k-means, softmax, LRU cache

System-design scenarios: enterprise document Q&A, real-time fraud detection, customer-support agent

Company-Type Playbook: what Big Tech, AI labs, startups, IT services, banks and healthcare typically emphasise

Project walk-through framework, HR and behavioural answers

7-day revision plan and last-minute checklist