AI Engineer (Claude / LLM Integration, Python) — South America, Fluent English — Short-Term

About the role We're building AI features into a production platform and need a Python developer who is strong with LLM APIs (Anthropic Claude). You'll build AI-powered automations: an agent that answers plain-English questions against live data using tool-calling, document/image extraction pipelines, and auto-drafting workflows — all wired into our existing FastAPI backend. ⚠️ Location requirement (please read): This role is open ONLY to candidates based in South America. Applicants outside of South America will not be considered, regardless of qualifications. You must also be fluent in written and spoken English. What you'll do - Integrate the Claude API (Anthropic) into a Python/FastAPI backend - Build an AI agent with tool-calling that connects to REST endpoints as tools and answers natural-language queries from live data - Build extraction pipelines: pull structured fields from web pages, receipts, and images (OCR-style parsing via LLM) - Implement auto-drafting workflows (e.g., generating outreach emails / structured records for human review) - Design reliable prompts, handle validation, retries, and error handling around LLM calls - Containerize and deploy AI services (Docker, AWS ECS) and collaborate via Git What we're looking for - Strong Python experience, ideally with FastAPI - Hands-on experience with LLM APIs — Anthropic Claude preferred (OpenAI/others acceptable if transferable) - Experience building tool-calling / function-calling agents (agent connects to APIs as tools) - Experience with structured data extraction from unstructured sources (documents, images, web pages) - Solid prompt engineering and an eye for reliability (validation, guardrails, retries) - Comfortable with Git-based workflows - Fluent English — clear written and verbal communication - Based in South America (required) - Self-directed, reliable, and strong at meeting deadlines Nice to have - Docker + AWS (ECS) deployment experience - Experience with agent frameworks and/or MCP (Model Context Protocol) - Workflow orchestration (Prefect, Celery, or similar) - Web scraping / content extraction (e.g., Firecrawl) experience - PostgreSQL / SQLAlchemy To apply In your proposal, please: (1) confirm your country of residence, (2) share links or code samples of LLM/AI work you've built — especially agents or extraction pipelines (GitHub welcome), and (3) tell us which LLM APIs you've worked with and what you built. Applications that don't confirm South America residency will be skipped.