Technical practice
How would you choose and evaluate a machine learning approach for an application?
What to coverSuggested practice: Explain how you would define the task, establish a baseline, and select evaluation measures that reflect the application's purpose. Discuss trade-offs such as model quality, inference cost, and the evidence needed before production use.
How would you design a production-ready AI application pipeline?
What to coverSuggested practice: Walk through data preparation, model inference, application integration, testing, and deployment in either a cloud or on-premises setting. Explain how you would check quality and handle failures without assuming a particular platform.
How would you integrate an AI solution into an existing workflow?
What to coverSuggested practice: Describe how you would identify the workflow's inputs, outputs, and dependencies with other teams. Explain how you would test that the integration works reliably and does not disrupt the existing system.
When would a generative AI model be appropriate for a solution?
What to coverSuggested practice: Compare a generative approach with a simpler machine learning approach for a clearly defined task. Address how you would evaluate output quality and manage unreliable responses before release.
How would you develop and assess a deep learning model using TensorFlow or PyTorch?
What to coverSuggested practice: Explain your choices of model architecture, training procedure, and validation method. Discuss how you would diagnose poor generalization and decide whether added model complexity is justified.
How would you develop and deploy a chatbot or conversational agent?
What to coverSuggested practice: Cover conversation scope, handling ambiguous requests, response evaluation, and integration with the application. Explain what you would test before deployment and how you would identify conversations the agent handles poorly.
How would you approach an image processing task?
What to coverSuggested practice: Describe how the intended output would guide preprocessing, technique selection, and evaluation. Discuss how differences in image quality could affect performance and what you would test.
How would you decide whether to use a cloud-based AI service?
What to coverSuggested practice: Compare a cloud service with an alternative implementation against the application's quality, integration, cost, and operational needs. Cloud-based AI services are listed as useful experience, so frame this as a design trade-off rather than an assumed requirement to use one.
How would you investigate and improve an AI system whose performance has declined?
What to coverSuggested practice: Explain how you would locate the problem across data, model behavior, and application integration. Describe the evidence you would gather, how you would evaluate a change, and how you would avoid worsening other outcomes.
Behavioural practice
Tell me about a time you coordinated with other teams to integrate a technical solution.
What to coverSuggested practice: Use a real situation to explain the shared goal and any conflicting needs. Describe your actions to align the teams and the resulting effect on the integration.
Describe a consequential technical decision you made as a subject-matter expert.
What to coverSuggested practice: State the decision, the options you considered, and how you used evidence and others' input. Explain your action, the result, and what you learned.
Tell me about a time you helped a team improve its performance.
What to coverSuggested practice: Set out the situation and the performance issue without attributing blame. Describe what you changed in how the team worked and the observable result.
Describe a problem you helped resolve that affected multiple teams.
What to coverSuggested practice: Explain the situation, why it crossed team boundaries, and how you identified who needed to act. Describe your contribution to a resolution and its outcome.
Tell me about a time you mentored a junior colleague.
What to coverSuggested practice: Describe the colleague's learning goal, the guidance or feedback you provided, and how you adapted it. Explain what changed as a result, using an outcome you can substantiate.