Technical practice
How would you design a RAG solution for a business question?
What to coverThe posting explicitly names Retrieval-Augmented Generation (RAG) and AI solutions that derive value from data. Suggested preparation: explain how you would select and prepare source material, retrieve relevant context, generate an answer, and evaluate whether it is supported by that context.
How would you diagnose a RAG system that gives plausible but unsupported answers?
What to coverThe role calls for Generative AI solutions that work at scale. Suggested preparation: distinguish retrieval failures from generation failures, then discuss test cases, answer-quality measures, and the trade-off between improving coverage and introducing irrelevant context.
When would you use an agent rather than a simpler AI workflow?
What to coverAgents and LangGraph or similar frameworks are named in the requirements. Suggested preparation: describe when multi-step orchestration adds value, how you would control state and tool use, and when a simpler workflow would be easier to operate.
How would you decide whether a problem needs Generative AI, another NLP method, or a different AI/ML approach?
What to coverThe posting asks for expertise in Generative AI, NLP, and other AI methods. Suggested preparation: compare approaches against the business objective, available data, evaluation criteria, and implementation complexity rather than assuming generation is always appropriate.
How would you turn a new AI research method into an implementable solution?
What to coverResearching new methodologies and translating advanced concepts into production-ready solutions are stated duties. Suggested preparation: describe how you would establish a baseline, test the method on relevant data, assess limitations, and decide whether the evidence justifies implementation.
How would you structure Python code for an AI/ML solution that others must maintain?
What to coverStrong Python proficiency and documentation of methods, tools, and processes are explicit requirements. Suggested preparation: discuss clear interfaces, reproducible experiments, testing, and documentation that lets collaborators understand and extend the work.
What would you consider when deploying a Generative AI solution in a cloud environment?
What to coverThe posting requires experience with solution deployment and cloud computing, without naming a provider or platform. Suggested preparation: explain deployment choices in terms of scalability, latency, cost, reliability, and how you would verify the deployed solution behaves as intended.
How would you evaluate whether an AI prototype is ready for large-scale use?
What to coverThe role emphasizes large-scale application and production-ready AI/ML solutions. Suggested preparation: distinguish promising research results from operational readiness by discussing quality across representative cases, failure handling, performance, and ongoing evaluation.
How would you frame and measure an AI use case across the energy value chain?
What to coverThe posting describes projects across the energy value chain and AI that supports data-driven business decisions and operations, but does not specify a particular project. Suggested preparation: show how you would clarify the decision to support, identify available data and stakeholders, and choose measures that connect model results to that purpose.
Behavioural practice
Tell me about a time you worked across disciplines or locations to deliver an AI-related outcome.
What to coverThe role involves multidisciplinary, multinational teams and global stakeholders. Suggested preparation: use situation, action, and result to show how you aligned different perspectives, handled communication, and contributed to a shared outcome.
Tell me about a time you helped others understand an AI capability or use case.
What to coverFacilitating training and increasing company-wide awareness of AI are stated responsibilities. Suggested preparation: describe the audience and its starting point, what you did to make the topic useful, and what changed as a result.
Tell me about a research collaboration with people outside your immediate team.
What to coverThe posting mentions collaboration with industry and academic partners. Suggested preparation: use situation, action, and result to explain how you shared knowledge, resolved differing priorities, and assessed the collaboration’s contribution without implying experience with a specific partner.
Tell me about a time you presented technical results to a customer or non-specialist stakeholder.
What to coverRegular presentations to customers and internal stakeholders, along with customer focus, are explicit parts of the role. Suggested preparation: explain the audience’s decision or concern, how you communicated evidence and limitations, and the resulting response or decision.
Tell me about a time new evidence changed your approach to an AI or data-science problem.
What to coverThe posting calls for staying current with AI developments, analytical judgment, and openness to learning. Suggested preparation: give the situation, what evidence prompted a change, the action you took, and the result or lesson.