Forward Deployed Engineer

Open Loop was co-founded by CEO, Dr. Jon Lensing, and COO, Christian Williams, with the vision to bring care anywhere. Our telehealth support solutions are thoughtfully designed to streamline and simplify go-to-market care delivery for companies offering meaningful virtual support to patients across an expansive array of specialties, in all 50 states.

Forward Deployed Engineering is how we build that platform with real clients instead of guessing. An FDE embeds with a client, works in the client's own codebase, and builds their integration against Open Loop's production API. Whatever the client needs that the platform cannot do yet is filed as a platform gap and built once, centrally, so every client inherits it.

The work product of an FDE is not the client's integration; it is the reusable asset that client forced into existence: a recipe others start from, an executable check behind every claim in our docs, a client manifest that makes the impact of any platform change computable, and the SDKs, agents and tooling that make the next deployment faster than the last.

You will write production code in two places, the client's system and ours. You will take an ask that is vague or over-promised and turn it into a scope you can defend. You will tell a client, and our own sales team, when something should not be committed to. And you will leave every deployment better instrumented than you found it, so the next client needs less of us than the last one did.

We are hiring for this role at the Senior, Staff and Senior Staff levels. Apply once; we will calibrate level during the process based on the scope you have carried, not on years alone. The section below describes what changes as scope grows.

• Build for clients, in their codebase, against our API. Embed with a client's engineering team and stand up their integration on Open Loop's Partner API: their app, their data, their identity, their legacy systems. Own the outcome through go-live and the weeks after it. Never wire around the platform: where the API falls short, file the gap with the endpoint it stands in for, and leave a dated, time-boxed workaround at most.

• Build with AI in the loop, for clients and for us. Ship the systems that let a client build on us with AI in the loop: SDKs and reference implementations, a CLI and machine-readable docs a client can hand to their own coding agent, and agentic workflows on our API such as intake, routing, support and refill journeys, with human review by design, evaluations before launch and monitoring after.

• Run deployments on the agent roster, and make it better. Use and extend the agents that run our deployments: the tenant readiness gate, conformance checks against a live tenant, seeding checkers, docs-freshness and superseding drafts, gap filing, client communication drafts. Every agent output is a draft a human approves, and you are accountable for knowing when the agent is wrong.

• Turn one client's need into every client's capability. Contribute additive recipes and checks to the shared cookbook, keep client manifests current, and bring the evidence that moves a recurring workaround into the platform roadmap instead of into another workaround. The measure is how much less of us the next client needs.

• Turn ambiguity into a plan. Take a client ask that is vague, over-promised or self-contradictory and produce a phased, defensible scope, with a clear statement of what we are not doing and what the API does not do yet.

• Never undo prior work. Checks lock the claims we publish; manifests tell you who depends on what; edits are additive or explicitly superseding with a compatibility window. A change that would break a live client stops and becomes a proposal,…