AI Engineer to Build a RAG-Powered Marketing Chatbot + Lead-Qualification Agents (Long-Term)

We're a B2B marketing agency running paid acquisition for ~30 SaaS and e-commerce clients. Our biggest bottleneck right now isn't traffic — it's what happens after the click. Leads land on our funnels, ask the same 40 questions ("do you work with my industry," "what's pricing like," "can you handle Klaviyo + Shopify," etc.), and our human SDRs are spending 6+ hours a day answering things that are already documented in our case studies, pricing decks, and onboarding PDFs. We need an AI engineer to build us a marketing chatbot that actually closes the gap between "interested visitor" and "booked discovery call." Not a glorified FAQ widget — a real RAG system over our internal marketing knowledge base (case studies, service decks, ICP docs, pricing logic, past proposals) that can answer questions in our brand voice and hand off cleanly to a human when the lead is qualified. On top of that, we want a layer of agents that do the follow-up work the chatbot starts: ● A qualification agent that scores the lead against our ICP and writes a short brief into HubSpot ● A scheduling agent that books the call directly into our SDR's calendar ● A nurture agent that drafts personalized follow-up emails if the lead drops off mid-conversation What we have: ● ~200 documents of marketing collateral (PDFs, Notion exports, Google Docs) ● HubSpot CRM, Calendly, and a Webflow site for embed ● A clear brand voice guide we'll share ● Strong opinions on what not to do (no generic "How can I help you today?" openers) What we need from you: ● Production experience with RAG (chunking strategy, embeddings, vector DB choice — we don't care if it's Pinecone, Qdrant, or Chroma, we care that you can defend the choice) ● Real chatbot deployments, ideally in a marketing or sales context — not just toy demos ● Comfort building multi-step agents (LangChain, LlamaIndex, or custom — open to discussion) ● You can explain why your RAG retrieves the right chunk 90% of the time and what you do about the other 10% Budget: Open. We'd rather pay for someone who's shipped this 3+ times than save 40% on someone learning on our dime. Expecting $4-8k for v1, with a retainer for iteration. To apply: Skip the template. Tell us about one RAG or agent system you shipped that's running in production today, what broke, what you fixed, and what you'd do differently. If you've built anything specifically for marketing or sales use cases, lead with that.