Agentic AI Engineer

Every day, millions of shipments move around the world on data that nobody is looking at in real time. Orders, ASNs, invoices, and tracking events flow between 3PLs, forwarders, brands, and retailers, then get flattened into a status that tells you something went wrong only after it already has.

SwiftB2B exists to change that. We are an AI operating system for global supply chains. One platform where 3PLs, freight forwarders, brands, retailers, and manufacturers can see every item, shipment, and exception as it happens. The platform does not stop at flagging a delay; tt works out what the delay costs, weighs the options, and starts the fix.

We built it on 15 years of SwiftAnt’s work in supply chain and EDI technology, and on a team that has spent careers in logistics, retail, and enterprise software. We are a Databricks partner and the platform runs on the lakehouse.

The “starts the fix” part is your job. Say a container is late into Jebel Ali. The agent needs to work out which orders are affected, what the customer’s SLA says, whether air freight on the top three SKUs is cheaper than the penalty, draft the message to the customer, and then stop and wait for a human to say yes. Every one of those steps has to be traceable afterwards, and the agent has to know when it does not know. We are looking for someone who has built enterprise-grade agents that do real work, who has seen them go wrong in production, and who designs the boundary before the capability.

What you will do

• Design and build the agents inside the product: exception triage, impact costing, corrective-action drafting, carrier and customer communication, document handling. Each one with an enumerated set of allowed actions and a human gate before anything irreversible.
• Own the reference architecture for agents on our stack: Claude Agent SDK, Databricks Mosaic AI Agent Framework, Model Serving, MCP servers over our internal tools and the semantic layer.
• Build the tools agents call: governed metric lookups, shipment and order APIs, rate and contract retrieval, EDI generation. Tools that return structured results and fail loudly.
• Write evaluation sets for every agent and run them before every release. Publish the pass rate. Try to break your own agent, and record what you found.
• Instrument everything: traces, tool calls, token cost, latency, escalation rate. When an agent acted, we can show why.
• Work with the solution engineers to get agents into the product behind proper identity and permissions, and with the quality team on the release gate for agents.
• Keep up with what the model providers ship each month and decide what belongs in the product and what is a demo.

What we are looking for

• You have shipped at least one agent to production that real users relied on, and you can describe how it failed and what you changed.
• Deep hands-on with at least one agent framework: Claude Agent SDK, LangGraph, Semantic Kernel, OpenAI Agents SDK, or the Mosaic AI Agent Framework. Experience building MCP servers is a strong plus.
• Strong Python. Comfortable with async, structured outputs, pydantic, tool schemas, retries, and idempotency.
• You have built evals and you know the difference between an eval that passes and a system that works.
• Clear thinking about boundaries: what an agent may do alone, what it must escalate, what it may never do, and how to test that adversarially.
• Experience with Databricks or Azure is a plus. Supply chain or logistics background is an advantage.
• A degree in Computer Science or a related field, or equivalent practical experience.

Not sure you tick every box?

Apply anyway. Nobody we have hired matched the list line for line. If you have built an agent that did something useful without supervision, and then watched it do something stupid, we absolutely want to talk to you. Send a link, a repo, or a write-up and tell us what the human gate looked like... that tells us more than a polished CV!

Equal opportunity

SwiftB2B hires on talent, skill and curiosity. We do not discriminate on the basis of race, religion, nationality, gender, sexual orientation, age, disability, or any other protected characteristic. If you need an adjustment at any stage of the process, tell us and we will make it happen.