Senior AI/LLM engineer to review a production multi-provider LLM system (Node.js)

We're a native AI front desk for hotels. Our backend (Node.js/Express) runs several LLM-powered surfaces: a guest-facing reply pipeline, a staff assistant with tool-use, background classifiers/extractors, and RAG over hotel knowledge. We route across multiple providers (Anthropic Claude, Google Gemini, Sarvam for Indic languages, OpenAI for embeddings) through a provider-abstraction and task-based router we recently built, with automatic fallback. We're looking for an experienced engineer to review this AI layer for correctness, cost-efficiency, reliability, and best practices before we scale. Deliverable: A written review with prioritized, actionable findings what's solid, what's risky, what to change, ranked by impact. A follow-up call to walk through it.