senior-prompt-engineer

Optimize LLM prompts by applying maintainable prompt patterns, few-shot prompting, and structured output constraints to improve accuracy and robustness in production AI systems. This ensures reliable, validated JSON responses rather than fragile demo-level outputs.

Set up repeatable LLM evaluation workflows to tune prompts and agents, measuring performance against defined criteria to catch and reduce regressions. This involves creating structured evaluation plans that span development and deployment scenarios.

Design agent architectures by planning orchestration patterns that ensure scalability, observability, and security for production deployment. This includes structuring agent interactions and monitoring constraints for robust system behavior.

Prompt engineering patterns directly improve RAG optimization by structuring retrieval-augmented generation prompts to enhance output accuracy and robustness. This applies to designing RAG-backed agents that must output validated structured data.

This approach works for end-to-end prompt architecture by spanning prompt design, agent orchestration, evaluation criteria, and production constraints like scalability, security, and monitoring. It turns raw prompt ideas into reliable, high-performance LLM behavior.