Technical practice
How would you lead the implementation of an LLM in an AI application?
What to coverExplain how you would define the application’s task, select an appropriate model and integrate it with the rest of the system. Address the trade-offs you would consider before calling the solution production-ready.
How would you evaluate an LLM-based application?
What to coverSuggested preparation: describe how you would test whether outputs meet the application’s goals and identify recurring failures. Explain how evaluation findings would inform changes to the model or application.
How would you design a production-ready cloud or on-premises AI pipeline?
What to coverWalk through how data and requests would move through the application, model and supporting systems. Suggested preparation: discuss how you would detect failures and maintain quality after deployment.
How would you decide whether to use a cloud AI service in a solution?
What to coverThe role includes designing AI solutions with cloud AI services. Compare a service-based approach with other deployment options in terms of integration, operational effort and fit for the application.
When would deep learning or neural networks be appropriate?
What to coverExplain what would justify their use rather than a simpler machine-learning approach. Image processing is listed as possible work; if relevant, describe how you would assess whether a neural-network approach suits that task.
How would you design an LLM-powered chatbot?
What to coverChatbots are listed as possible work. Suggested preparation: explain how you would define the chatbot’s scope, integrate it into an application and test responses that fall outside that scope.
How would you choose among linear regression, logistic regression and decision trees?
What to coverRelate each method to a suitable prediction task and explain the assumptions or interpretability trade-offs that affect your choice. Describe how you would check whether the selected method performs adequately.
When would you use clustering, and how would you assess its results?
What to coverExplain how clustering differs from the listed supervised methods and what makes a grouping useful for a particular problem. Address how data preparation and the choice of features can change the result.
How would you clean, transform and normalize data before modelling?
What to coverDescribe how you would identify data-quality issues and choose transformations suited to the model and data. Explain how you would check that preprocessing preserves integrity and is applied consistently.
How would you use Tableau or Power BI to communicate an analysis?
What to coverThe posting asks for experience with either tool. Explain how you would choose a visualization that makes a statistical or model result understandable, including limitations the audience should know.
Behavioural practice
Tell me about a time you led a team through a technical decision.
What to coverUse a real situation, the decision you were responsible for and the actions you took to reach it. Explain the result and how you handled differing technical views.
Describe a time you worked across functions to integrate a technology into an application.
What to coverSet out the teams’ different needs, what you did to coordinate the integration and the outcome. Keep the example focused on your own contribution rather than attributing the whole result to yourself.
Tell me about a problem you helped solve across multiple teams.
What to coverDescribe the situation, why the issue crossed team boundaries and the actions you took to resolve it. State the result and how you confirmed the solution addressed the shared problem.
Give an example of how you supported a team’s performance as a subject-matter expert.
What to coverExplain the team’s challenge, how you shared expertise or removed an obstacle and what changed as a result. Distinguish your actions from the team’s collective work.
Describe a time you researched a new AI technique and decided whether to apply it.
What to coverExplain the problem, how you evaluated the technique against alternatives and what you recommended or implemented. Include the outcome, including what you learned if the technique was not adopted.