We are seeking an experienced AI engineer to develop a lightweight backend service that extracts structured data from unstructured text using a large language model (LLM). The objective is to build a reliable API that converts raw text inputs (such as inbound emails, support tickets, or short documents) into validated, schema-compliant JSON outputs. This project is a focused prototype intended for internal workflow automation. This is a backend-only engagement. The emphasis is on structured output enforcement, reliability, and clean architecture rather than UI development. Responsibilities: - Develop a Python-based API (FastAPI preferred) - Integrate OpenAI or a comparable LLM provider - Implement structured output using function/tool calling or JSON schema enforcement - Validate LLM responses against a predefined schema - Implement retry logic for malformed or invalid outputs - Add structured logging for traceability and debugging - Ensure modular code organization and separation of concerns - Provide concise setup and usage documentation Technical Requirements: - Strong experience with Python and API development - Prior experience integrating LLM APIs (OpenAI, Anthropic, etc.) - Familiarity with structured output patterns and schema validation (e.g., Pydantic) - Understanding of prompt design for consistent, deterministic outputs - Clean dependency management and environment-based configuration Deliverables: - Functional API endpoint that accepts raw text input and returns validated JSON output - Schema validation layer - Source code Timeline: 2–3 business days To confirm you have carefully reviewed this posting, please begin your proposal with the word Banana and clearly include your proposed budget within your cover letter. Please include a brief summary of relevant experience with LLM integrations and structured output systems.