Founding Forward Deployed Engineer

A booking changes. A document arrives late. Someone has to work out what it means, decide what happens next and make sure the right system reflects it. We're building AI that can help carry that work, with people in control of the decisions that matter. As a founding Forward Deployed Engineer, you'll work directly with operators and turn that understanding into production software. You'll follow a problem from the first conversation through the integration, the evaluation and the first real exception. Your decisions will shape both our product and how we build. What you'll build with us
- Sit with customer teams, trace their workflows and find the smallest useful thing we can put into production.
- Build AI workflows, integrations and interfaces that work with imperfect documents, changing requirements and existing systems.
- Design evaluations, approval boundaries, monitoring and recovery paths so the software can earn an operator’s trust.
- Stay close after launch: watch how the system is used, resolve failures and turn what you learn into reusable product capabilities.
- Help establish our engineering standards and make thoughtful calls about what to build, buy or leave out. Who we're looking for You have shipped AI that people use. You can walk us through a workflow in production: what the model does, which tools it can call, how state or memory is handled, how you evaluate it and what happens when it fails. You have stayed to fix the awkward cases. AI coding tools are part of your working day. You use tools such as Codex or Claude Code to build and investigate software. You can enter an unfamiliar codebase, trace a failure across APIs and data, and explain the code you accept. Python or TypeScript experience is useful. You go looking for the operational truth. You will sit with an operator, inspect the source documents and find out why the spreadsheet still exists. Freight experience helps, but so does evidence that you have learned an unfamiliar operation well enough to improve it. You can build before the playbook exists. You have delivered with lean resources and incomplete requirements. You can choose a small useful release, decide what needs your own hands and explain which shortcuts will need attention later. Show us where your decisions affected delivery. You connect engineering to the business. You want to know who needs the workflow, what the current problem costs and whether anyone uses the fix. You can challenge an impressive feature that will not help the customer and propose something worth shipping instead. How we work You choose useful work without waiting for a task queue. When a route closes, you investigate another, bring a recommendation and follow through. You ask for help before a blocked decision becomes a missed commitment. You catch weak assumptions, generic reasoning and fabricated details. You check AI outputs before passing them to a colleague or customer. The tools can do more of the work; you still own its quality. You tell the truth early. You flag doubts while there is still time to act. You can disagree with a founder, admit uncertainty and change your mind when the evidence changes. Manufactured agreement helps nobody. You take endurance seriously. You want to do demanding work well over time. You protect the habits and recovery that sustain your judgement, make realistic commitments and raise unsustainable workloads before exhaustion makes the decision for you. Your first chapter Get one real workflow into people's hands. Understand the source records, build a thin end-to-end version, agree how success will be checked and improve it with the people using it. Why join at the beginning? You'll help decide what we build, which customers we pursue and what good work looks like here. Expect incomplete systems, direct customer contact and decisions whose consequences you can see. We want people who enjoy making something work while the answers are still taking shape. Early-stage experience carries more weight here than a corporate title. What matters is evidence that you can build, learn and take responsibility when resources are lean. If you developed that judgement inside a larger organisation or took an unusual route, show us the work. YOUR NEXT CHAPTER Let's build something useful. Show us an AI workflow people use, what you personally built and a failure you had to fix. Tell us how AI tools fit into your work and where you had to challenge their output.