AI Content Pipeline Engineer (Python/LLM APIs)

I am mobile trivia game powered by a structured knowledge graph (hierarchical categories, subcategories, geographic/regional dimensions, and specific subject entities). Our master taxonomy is already established in a structured spreadsheet covering approximately 2,500 priority nodes. Looking for an AI Automation & Data Engineer to design, execute, and deliver a production-grade automated batch-generation and multi-pass validation pipeline. You will generate 10,000 verified multiple-choice questions and ~5,000 paired image assets, writing all validated data directly into a Supabase (PostgreSQL) database. Technical Stack & Qualifications Languages & Database: Python (AsyncIO, Pydantic, Pandas), PostgreSQL / Supabase. LLM Orchestration: Proven experience Anthropic Batch APIs, strict JSON schema validation (via Pydantic / Instructor), and search-grounded models. Vector Embeddings: Semantic similarity scoring and deduplication (e.g., text-embedding-3-small). Image Generation & Multimodal QA: Experience orchestrating visual APIs (FLUX.1, Midjourney API, or Ideogram) coupled with automated vision QA workflows.