Forward Deployed Engineer

Blank Metal•Minneapolis, MN

14d•Hybrid

About The Position

Enterprises come to Blank Metal to get production AI live. That work is led by senior people who can sit across from a client's executives, diagnose the real problem, and then go build the thing that fixes it. As a Forward Deployed Engineer you'll turn advisory conversations into shipped software, closing the gap between a promising demo and a system that runs in production, where most enterprise AI initiatives stall. You embed with a client, earn trust at the executive level, and stay hands-on from the first scoping session through the live deployment. You are equal parts technical leader and trusted advisor: you shape what gets built, you build it with a lean team of AI engineers, and you own whether it actually works in production. The playbook for this practice is still being written, and you will write a lot of it: how we embed, how we prove value, and what we bring back from each engagement. Forward deployed means what it says. Expect roughly 25 to 40% travel to client sites, heavier during kickoffs and go-lives. The rest of the time you work wherever you work best. This is a senior, high-autonomy role with real ownership: you carry a small number of accounts deeply, and you are measured on outcomes clients can see.

Requirements

Typically 7+ years building and shipping software, including real time as a technical leader who executives actually trusted
A track record of owning production systems end-to-end, from first commit through live operation
You have shipped LLM-powered systems to production: agents, retrieval, evals, and the judgment to know when each is the wrong tool
Strong hands-on skills in Python and TypeScript or JavaScript, and comfort picking up whatever the client's stack demands
You have built or maintained eval frameworks that showed whether a system actually worked
You can hold your own with a CTO in the morning and be in the codebase in the afternoon
At home with ambiguity and a small team: you would rather figure it out than wait for a playbook
A straightforward communicator who explains hard technical calls in plain language
Low ego, high cooperation: you balance competing priorities with integrity and make the people around you better
You have been a founder, CTO, or first engineer building a product or startup and know what it costs to get something live.

Nice To Haves

Depth in healthcare or private equity, where much of our current work lives

Responsibilities

Partner with client executives to find the highest-value AI opportunity, then frame it in the client's own language
Lead scoping and proof of concept work that de-risks the build before a full engagement starts
Shape the technical approach and scope alongside our growth team; they own the close, you make sure what's promised is what can actually be built
Deploy alongside a lean team of AI engineers and own the architecture, the hard calls, and the code that ships
Deliver the concrete artifacts clients run in production: MCP servers, sub-agents, agent skills, and the pipelines around them
Work across the full stack and the model layer
Navigate real enterprise environments: the client's IT systems, security constraints, and the stakeholders who own them
Stand behind results with observability and evals the client can check for themselves
Win over the people who use the system day to day; a deployment succeeds when they rely on it
Carry accounts end-to-end: some weeks are executive workshops, some weeks are debugging and agentic orchestration
Bring patterns back to the team so the next engagement starts further ahead