I will build an ai chatbot with rag over your documents

Most document chatbots fail the same way: they answer confidently and wrongly, and nobody notices until a customer does.

I build retrieval that is measured, not assumed. Hybrid search combining keyword and vector similarity, chunking chosen by testing rather than guessing, and a set of evaluation questions so you can see the quality instead of trusting a demo.

WHAT YOU GET

A working chatbot over your documents (PDF, Word, web pages or database), deployed and documented, with an evaluation set so quality is a number and not a feeling.

WHY ME

I have shipped LLM integrations in production across four products on OpenAI, Anthropic and Google Gemini. My open-source project runs hybrid RAG over 9,125 chunks with retrieval graded automatically in CI: a 32-question gold set that fails the build if quality drops. You can read the code: github.com/RodrigoFK06/PrecioVivo

Two Anthropic certifications. Founder of a software consultancy: 50+ projects for 45+ clients since 2020.

BEFORE YOU ORDER

Message me with what your documents are and what people will ask. I will tell you honestly whether RAG is the right answer for your case, and if it is not, I will say so.