I will add llm features to your python backend with cited rag and cost control

You want an LLM feature that gives grounded answers, fails safely and doesn't surprise you on the bill. I build that into existing Python backends.

What I build: retrieval (RAG) with citations, so every answer points to its source; structured outputs, validated with pydantic before anything is saved or acted on; agent steps that wait for a person's approval before they change data; model routing and caching to keep the cost per request in check; and an evaluation set, so quality is measured rather than eyeballed.

How I work: I read your codebase before changing it, write the failing test first, and walk you through the feature on your own data before handing it over. Background: production Python (Django REST, pytest, queue workers on AWS) on a platform that already ships LLM features, after years of remote work for teams in New Zealand, Singapore and Australia.

What you get: the feature in your repo with tests, a short design note and the evaluation results.

Message me first with your stack and what the feature should do.