Product Manager- Logistics

About Tensec Tensec is Vertical AI for cross-border trade. Trade is two flows, money and goods, and the trading companies that move them: freight forwarders, shipping lines, customs brokers, 4PLs — still run both on systems built for large corporations and a lot of manual work. Tensec is the agentic AI layer those companies use to move both flows for their importer and exporter clients: FX and payments today, quoting and booking next, trade finance to follow. Cross-border B2B is a $50 trillion market. We're building the version all global trading companies choose to use. We started in the US, Mexico, and Brazil and are expanding globally, onboarding enterprise trading companies whose trade flows run to billions of dollars a year across every major corridor. Founded in 2023 by a team that has built and exited startups and run global teams at Meta, PayPal, Credit Karma, Goldman Sachs, Visa, Mastercard, and American Express. Tensec is backed by world-class investors such as Quiet Capital, Coinbase Ventures, Costanoa Ventures, Renegade Partners, Clocktower, and WillowTree. Your mission Getting freight across a border still runs on email and on the memory of whoever is handling it. Someone prices the request by hand against contracts and this month's surcharges, keys much of it in again to make the booking, and watches for what changes while the shipment moves. Most of what makes any of it go well lives in that person's experience rather than in a system. You are joining us to change this. You will own the AI for that run of work: the request, the quote, the booking, and keeping the shipment honest while it moves. Those stages are your first year. They are also what everything downstream stands on, because the commercial and financial side of a forwarder only works as well as the shipment file underneath it. This role has a forward deployment component, and we mean it literally. You will sit with our clients' pricing and operations teams, learn how the work actually happens, and design workflows that fit their process. Engineers build the connections into their systems, while you. You decide what moves between them, which parts of the work an agent takes, and where a person stays in the loop. We are early in this product line. The first clients are how we find out what can be generalized and productized, and you will be the person making that call. If that is the interesting part of the job to you, this is your role. Outcomes — your first 12–18 months Get an agent into daily use inside a client's team within your first two quarters. One workflow, doing work a person did by hand the week before. Establish the baseline nobody has today. Pick the few numbers that matter, measure them honestly before you build anything, and prove the improvement against them rather than asserting it. Stop the handoffs losing information. What one stage decides should reach the next without being retyped, and what changes mid-shipment should reach the people it affects without someone remembering to forward an email. Get the judgment out of people's heads and into the product. The calls experienced operators make without thinking are the product, and they need to survive the person who knows them leaving. Set the accuracy bar for an AI system where a wrong number costs real money. Define what good means, build the eval set, design the confirmation step, and decide what an agent is never permitted to send unsupervised. Turn client-specific work into product. For every workflow, make the call: bespoke, configurable, or core, and be able to defend it. Bring the second and third clients onto the same product with materially less custom work than the first. Also on your plate Requirements gathering as a craft. Running sessions with pricing, operations and finance teams, and leaving with a process map they correct rather than politely accept. Writing the specs and flows engineers build from, and staying close enough to the build to answer questions in hours rather than sprints. Putting a working prototype in front of a client instead of a slide. You will build those yourself with AI tools. Being the person the client's pricing lead calls. You own that relationship rather than handing it to an account manager. Working through the hard cases rather than around them. The requests that do not fit the standard path, and everything that changes after a shipment is confirmed, are the workflow rather than the edge case. Partnering with the freight expertise we already have in house, and helping shape what we charge for and how. What you bring 5+ years building or delivering software for operational teams, including at least one 0 to 1 you took from nothing to real usage. Product management is the most common path into this role, but we will read implementation leads, solutions leads and operators who moved into product with equal interest. We care more about trajectory than title. Experience in a domain where the pricing and operational data is genuinely messy. You ask about the data model before you ask about the screen, and you know what a business rule looks like after ten years of being patched by hand. You can read a process and see the data underneath it. Given two systems and a handoff between them, you can work out what each one owns, where the truth lives, what goes stale, and what breaks when a person is the integration. Judgment about where AI belongs and where it does not. You can look at an operational workflow and say which parts an agent should own, which parts need a person, and which parts should stay deterministic because control matters more than cleverness. Hands-on with AI, in the way a product person is. You have built with LLMs yourself, even if only in prototypes, you are comfortable with prompting, retrieval and evaluation, and you can tell a demo that works from a system that works. You do not need to write production code. You are good in a room with a customer. You can run a working session with people who have been burned by software before, and leave with them on your side. You can hold your own with engineers without doing their job. You ask good questions about what a system can and cannot give you, and you know when "that's hard" means hard and when it means tedious. Working Spanish is a plus. Our client pricing teams work in Spanish, and the signal you need is often in what they say to each other rather than in what gets translated for you. Not a gate if everything else is there. Based in Miami, NYC, or Palo Alto, and willing to travel on site with clients in the US and Latin America. How you work You go and look. A day shadowing a pricer beats a well-run interview, and you would rather do both. High agency, high urgency. You decide and move rather than waiting for a room to converge. You say no to some of the wish list items. Clients will give you twenty ideas. You ship the one that moves the number, and you can explain that choice to their CEO.You write. Specs, decisions, and what you learned this week, so the rest of the company can build on it. You hold a date. When something is going to slip you say so early and propose a smaller version rather than a later one. You would rather ship a narrow thing that works than a broad thing that demos.