Forward Deployed AI Engineer

## About the Role Forward deployed means you're embedded. You'll sit inside a client's team as the engineer who takes an ambiguous business problem and turns it into a working, maintainable production system. You'll move between engagements as projects ship, so you'll see more codebases, stacks, and founding teams in a year than most engineers see in five. You're also the technical face of Arkanis in those rooms, which means the quality of your work and your judgment both matter. This role is based in Kitchener-Waterloo. You'll work from our Kitchener office up to three days a week, with the rest remote, plus time on client sites during active engagements. We're around to pair on hard problems and back you up however you need it. This is a role we expect to grow someone into: with strong fundamentals and real curiosity about applied AI, we'll teach you the production ML and agent engineering on the job, with senior ICs to learn from. ## What You'll Work On No one person does all of this. Think of it as the range our engagements pull from, the kind of work you'll get to lean into and grow across over time: - Setting up the testing backbone that turns 'move fast and break things' into 'move fast and keep them working.' - Building the testing and CI that let us ship AI features fast and sleep at night. - Designing thoughtful evals and benchmarks, model guardrails, experiment design, observability, and A/B tests, so we actually know whether a feature is working. - Prompt tuning and shaping model behavior until the results are reliable. - Improving stability and performance across the board. ## About the Company We're a production AI systems shop. Growth-stage companies bring us in when they have a real engineering problem and not enough senior bandwidth to solve it well: data infrastructure that's buckling, an LLM feature that breaks the moment real users touch it, a matching engine that needs to actually work. We embed with their teams, ship the system, and harden it for the long haul. Our work spans LLM agents, data pipelines, and applied ML across a rotating set of clients. ## Who You'll Work With Neither founder came up the standard way. Both started in the hard sciences and taught themselves computer science and machine learning from there. They worked with the first generation of transformer models, shipped some of the earliest commercial LLM applications running inside Fortune 500 companies, led engineering teams at unicorn AI startups, and put research into production at world-class labs including Microsoft Research and DARPA. None of it came with an Ivy League stamp or a FAANG badge. We earned it by shipping, and that's the only resume line we actually read, and the only thing we test for. No leetcode theater in our interviews, just real systems and whether you can build and reason about them.