01 · Python · LIVE Thrifty Router
An intelligent proxy for Google Cloud Vertex AI that routes prompts to the cheapest capable Gemini tier (lite, standard, pro) via semantic similarity, zero-shot classification, or opportunistic cascade with multi-factor verification — cutting API spend 61% at Pro-level judge scores, or 97% for a 2% quality drop, on a 300-prompt benchmark.
PYTHON · FASTAPI · VERTEX AI · GEMINI · LLM ROUTING · VECTOR SEARCH · CLOUD RUN · FIRESTORE