Enterprise RAG and document-intelligence pipelines I built as Head of AI at Rivery (acquired by Boomi). Multi-stage indexing with AI-native search engines (Perplexity, Tavily) and smart chunking, hybrid retrieval with Cohere reranking and OpenAI embeddings, multi-LLM features across OpenAI, Anthropic Claude and Cohere, and Python evaluation pipelines covered by unit, integration and E2E tests. I can build the same for you: a RAG chatbot or AI search over your own documents, with measured accuracy.