1+ year of hands-on experience with a major cloud platform (Azure, AWS, or GCP), including working proficiency in observability tools (Application Insights, CloudWatch, Datadog, Splunk, or similar) to debug issues from logs and metrics, and the ability to write basic log queries (KQL, SQL, or similar). AI Realism and Eval Fluency: Understands the difference between deterministic logic and probabilistic generation; designs guardrails for hallucination, bias, and drift; uses evaluation harnesses before launch and monitors drift after, with a kill-switch mentality-and knows when not to use AI.