AI Agent Developer Needed - Action-Taking Support Assistant (LangChain, Python)

We run a customer-support operation and want to move past the standard "chatbot that just answers FAQs." What we actually need is an AI agent that can *do* things - pull real information from our systems, take the right action, and recognize when a situation should go to a person instead. If you've built tool-calling agents that operate on live internal data rather than canned responses, this is for you. The core behavior we're after: - Connect to our internal tools to retrieve order details, run status checks, and open support tickets - Reason about the conversation and pick the correct tool for each situation - Route the conversation to a human when the request is high-stakes or falls outside what it should handle - Stay dependable — it must never invent an action or fire on incomplete/incorrect data Scope of work: - Develop the agent on LangChain using a tool-calling design - Connect 3-4 of our internal APIs as callable tools (order lookup, status check, ticket creation) - Layer in validation and guardrails so every action is deliberate and safe - Set up a clean handoff path to human support for out-of-scope cases - Make sure the agent runs predictably once it's live What we're looking for: - Proven track record building tool-calling agents (not basic chatbots) - Solid command of LangChain and Python - Comfortable wiring agents into internal tools and APIs - Knows how to enforce guardrails, validation, and reliable behavior - Writes clean, documented code Nice to have: - Built human-in-the-loop handoff flows before - Background in support or service automation Already built something like this? Share a quick example or portfolio link with your proposal.