Defend

40% lower cloud cost on GCP

Cost discipline is an architect trait, and they’ll ask how it carries over to GenAI.

On your resume

Reduced cloud-processing cost by 40% and manual effort by 80% through pipeline optimization, serverless automation, and operational standardization.

Cross-questions

2
A strong answer covers
  • Specific levers: BigQuery partitioning and clustering, no SELECT *, right-sized Dataflow workers, consolidated Airflow DAGs, serverless for spiky jobs.
  • Measured on billing over a fixed period.
Avoid

“Optimization” without one concrete lever.

A strong answer covers
  • Token budgets, caching, routing to smaller models, prompt trimming, batch APIs, and self-hosting only at steady high volume.
Avoid

Missing the link between your data work and LLM costs.

Your numbers

Interviewers probe numbers first. Write exact values so you never have to improvise.

Next claim