Question bank

Interview questions, with guidance on how to answer

834 questions across 64 roles. Open one to see what a strong answer includes and common slips, then drill it out loud with Ava. New to answer structure? Start with the learning hub.

834 questions.

TechnicalSeniorMachine Learning Engineer
If you were setting up the machine learning platform for a team of ten scientists, what would you build first and what would you buy?
Tests: MLOps & deployment
BehaviouralSeniorMachine Learning Engineer
Tell me about a time you chose not to use machine learning, or chose a much simpler model, and why.
Tests: Pragmatism & ownership
TechnicalAny levelMachine Learning Engineer
Beyond accuracy, what do you monitor for a model running in production?
Tests: Model monitoring & reliability
TechnicalAny levelMachine Learning Engineer
How do you version data and features so you can reproduce a model from six months ago?
Tests: Data & feature pipelines
TechnicalEntry levelAI Engineer
In simple terms, what is retrieval augmented generation, and why would you use it instead of just asking a language model directly?
Tests: Retrieval augmented generation
TechnicalEntry levelAI Engineer
A chatbot you built keeps giving long rambling answers. What would you change first?
Tests: Prompt & context engineering
TechnicalEntry levelAI Engineer
What is prompt injection, and how might it affect an assistant that reads customer emails?
Tests: AI safety & security
TechnicalMid levelAI Engineer
Your document search assistant often answers using the wrong document. How would you debug and improve retrieval?
Tests: Retrieval augmented generation
TechnicalMid levelAI Engineer
How do you evaluate whether a change to a prompt or model actually made your application better?
Tests: LLM evaluation
TechnicalMid levelAI Engineer
Tell me about an agent or tool-using workflow you built. How did you stop it from looping or taking the wrong action?
Tests: Agents & tool use
TechnicalMid levelAI Engineer
Your app's language model costs are growing faster than usage. What levers would you pull?
Tests: Production reliability & cost
TechnicalSeniorAI Engineer
When would you build a fixed workflow with a model at certain steps, versus giving an agent freedom to plan its own steps?
Tests: Agents & tool use
System designSeniorAI Engineer
How would you design an AI assistant that can take actions in company systems, like issuing refunds, so it stays safe?
Tests: AI safety & security
BehaviouralSeniorAI Engineer
Tell me about a time you had to tell a stakeholder that an AI feature wasn't reliable enough to launch. How did you handle it?
Tests: Product & user focus
BehaviouralAny levelAI Engineer
Models and tools in this field change every few months. How do you decide which new things are worth adopting?
Tests: Learning agility
TechnicalAny levelAI Engineer
How do you handle latency in a chat or voice product where users expect a reply almost immediately?
Tests: Production reliability & cost
CaseEntry levelAutomation Engineer
How would you decide whether a manual process is a good candidate for automation?
Tests: Process analysis & discovery
TechnicalEntry levelAutomation Engineer
Tell me about a small script or automation you've built, even just for yourself. What did it do and how did you test it?
Tests: Scripting & programming
TechnicalEntry levelAutomation Engineer
An automation that copies invoices into the finance system fails halfway through a batch. What should happen next?
Tests: Error handling & resilience
TechnicalMid levelAutomation Engineer
When would you use screen-based robotic process automation instead of connecting through an API, and what are the risks?
Tests: RPA & workflow tool development