Akarshan User

I design, test and optimize reusable prompts, AI playbooks and structured LLM workflows for business problems, with active independent projects spanning prompt libraries, evaluation systems and an AI customer support assistant.

My Prompt Engineering Portfolio documents 50–100 reusable prompts across ChatGPT, Claude and Gemini, including structured JSON outputs, classification, extraction, summarization, guardrails and prompt-injection handling. I maintain version-controlled prompt iterations, outputs, evaluation criteria and optimization rationale.

I built an LLM Prompt Evaluation System that compares prompts, models and workflows against shared test cases using accuracy, relevance, completeness, consistency, format compliance and hallucination-risk metrics.

Before moving into AI workflow design, I inspected and validated over 1,000 precision components at BWT Alpine F1 Team and led 18 staff at Forest of Arden Hotel and Country Club. I bring disciplined documentation, root-cause analysis, process improvement and team-training experience to building reliable AI systems.