LLM Architecture and Orchestration Expert

Most llm architecture and orchestration work breaks on exactly the cases teams skip until production: context construction quality and relevance filtering strategy, and prompt-tool-retrieval contract boundaries and error propagation.

What LLM does:

• context construction quality and relevance filtering strategy
• prompt-tool-retrieval contract boundaries and error propagation
• structured output constraints and downstream parsing robustness
• fallback/degradation strategy for model/tool/retrieval failures
• eval design: scenario coverage, success metrics, and regression detection
• latency/cost budget alignment with product requirements
• orchestration complexity versus debuggability and maintainability

What you get:

• SOUL.md — LLM's identity and working methodology
• llm-architect.md — the full persona instruction file
• MEMORY.md — session-persistent context template

Install:

# 1. Place in your project
cp llm-architect.md .claude/personas/LLM.md

# 2. Add to CLAUDE.md
echo "## Active Persona\nLLM handles llm architecture and orchestration work. See: .claude/personas/LLM.md" >> CLAUDE.md

# 3. Call by name in Claude Code
# LLM, [your task here]

LLM maps the problem space before writing a line, validates success and failure paths, and reports residual risk honestly. Use it when production discipline matters more than speed.