Supply Chain AI Agent Engineer Consultant

About DarkchainToday's supply chains are complex systems held back by disconnected workflows, bad data, and slow human decision-making. The cost shows up as late shipments, broken OTIF performance, blown forecasts, and an inability to respond when something goes wrong, which in a real supply chain is always.Darkchain is building the AI agents that fix this at the root. We treat AI as the machine which an operator uses: the layer that actually does the work that humans used to do by clicking between systems, chasing exceptions, and patching bad data. Our agents monitor fulfillment workflows 24/7 and catch silent failures before they hit the customer. They connect siloed systems end-to-end so an order's journey is finally visible and actionable. And they act as tireless guardians of master data, continuously detecting and correcting the errors that quietly poison every downstream forecast and plan.We are not building another chatbot. We are building the operational substrate that makes a supply chain genuinely resilient. Agents that observe, decide, and act, with humans in the loop where it matters and out of the loop where it doesn't.The RoleWe are hiring an Agent Creator. You will design the machines. Ignore the Job title above. You will be an owner with equity.Your core job is to decompose a human operator's work, whether that is a warehouse coordinator, a planner, or a master data steward, into discrete decisions and actions, and then build an AI agent that performs those decisions and actions reliably in production. You will sit close to real operators, watch what they actually do (not what their job description says they do), and translate that into an agent that can do it at a fraction of the cost, around the clock, without getting tired or distracted.This is a founding role. You will own how our agents are designed end-to-end: what they perceive, what tools they have, how they decide, how we know they're working, and how they get better over time. You will ship the first agents that real customers depend on, and you will shape how Darkchain builds every agent that comes after.What You'll DoEmbed with our first design partners and reverse-engineer the workflows their humans currently run, whether that is fulfillment exceptions, OTIF orchestration, master data hygiene, or whatever the highest-pain workflow turns out to be.Break those workflows down into the discrete decisions and actions an agent needs to take, the systems it needs to read from and write to, and the failure modes it has to handle.Design and build the agents themselves: the prompts and decision logic, the tools and integrations into WMS, TMS, ERP, EDI and whatever else the workflow touches, and the eval suites that tell us when an agent is regressing.Run the agents in production. Watch the traces. Catch the failure modes nobody anticipated. Fix them. Make the agent boringly reliable.Help shape the platform underneath the agents, including what reusable patterns, tools, and infrastructure should exist so the second, tenth, and hundredth agent are faster to build than the first.QualificationsYou can describe systems with precision. Agents are built out of language. The person who can write a clear, unambiguous description of how a workflow actually works is the person who can build an agent that runs it. We are explicitly open to authors, poets, journalists, technical writers, analysts, lawyers, philosophers, librarians, and screenwriters. Anyone whose craft is taking something tangled and rendering it in language a reader (or a machine) cannot misinterpret.You think in logical systems. You can take a messy human workflow and break it cleanly into the discrete decisions and actions a machine could perform. You see the dependencies between parts. You notice when one rule quietly contradicts another. You can spot which steps are actually hard versus which steps just look hard.You are curious about supply chain, or you already know it. Prior experience in supply chain, logistics, warehousing, planning, or operations is a strong plus but not a requirement. What is required is a willingness to spend real time learning how a warehouse actually runs and to take the domain seriously instead of treating it as a generic backdrop for AI.You are self-directed. This is a founding role at a small company. Nobody is going to hand you a ticket queue. You decide what to build next based on what will most move the company forward.What We Are Explicitly Not Looking ForWe are not looking for a coder only. We are not looking for a computer scientist only. Programming is not the bottleneck of this job. Designing the agent is. If your strongest skill is writing production code, you are probably not the person we are hiring first.A CS degree is not required. No degree is required at all. Skill is something you demonstrate by what you can do, not by what credential you hold. If you came up through the humanities, through operations, through a trade, through self-study, or through any other non-traditional path, you are exactly the kind of person we want to hear from.We are not looking for a generalist ML researcher who wants to fine-tune models from scratch. That is not the work. The work is designing the machine that replaces the operator.We are not looking for someone who confuses a chatbot demo with an agent. The job is to design systems that observe, decide, and act in the real world, not to write clever prompts for a Q&A box.How to ApplyApply. Just try. We will use our own tests to qualify you through the rounds.