Real production experience building on large language models and generative AI: agents, retrieval, evaluation frameworks, prompt management, guardrails, and the harder operational realities of running AI in production (rate limits, non-determinism, provider outages, schema drift, cost and latency). Enough depth in distributed backend systems and cloud-native architectures (AWS, Azure, or GCP) to credibly review what your teams ship: services, async processing, event-driven pipelines, and content and data systems at scale.