Betterup Staff Forward Deployed Engineer leads technical customer deployments and builds production software for enterprise AI systems. The role involves managing technical relationships, writing code, and raising engineering standards.
Responsibilities
AI is changing how every person in the world works. It's a new Industrial Revolution, and almost all of the attention is on the technology. Far less is going to the people who have to change how they work to use it.
We think helping people through that change is one of the most important problems in the world right now. Even the most technically advanced companies are struggling to get their workforces productive with AI, and our research shows much of the gap comes down to mindset and behavior, which can be coached. Coaching used to be something a company could afford only for its top leaders. With AI and human coaches working together, we can reach everyone, and the most iconic companies in the world are asking us to build that with them.
Our largest customers no longer decide on the demo. They decide on what happens after: whether their HR systems connect, whether provisioning works, whether coaching shows up inside the tools their people already have open, whether our AI behaves the way their risk team needs it to. Forward Deployment is the function that owns that work.
This role sets the technical architecture for Forward Deployment. Our largest customers are asking for the same few things in different words: the coach inside their own assistant, their data in their own lake, an AI deployment their governance committee will sign. Today we solve each of those one account at a time. You decide what the shared pattern is, build the first version on the hardest account, and make it what every squad uses and what the core platform eventually absorbs.
The accounts are specific. One customer wants our coach answering inside the agent platform it is rolling out to its whole workforce. Another is building its own experience layer on top of ours and needs an architecture where each experience is a declarative object instead of a custom build. A third signed, and then its AI governance committee asked how its data is hosted, whether it trains a model, and what happens under a legal hold.
You'll do this as a hands-on engineer, with squads of engineers and PMs around you, and you'll be the engineer we put in front of a customer's CTO, chief architect, or CISO when the technical answer decides the deal.
Set the architecture on the hardest accounts in the book. You walk into a customer whose environment we don't know yet and leave with a design their architects accept, our engineers can build, and core engineering can live with.
Decide whether we build, configure, or refuse. Some asks are a week of configuration, some are a real build, and some would fork the product. You make that call early, in writing, and you defend it to the customer and to our own leadership.
Own the pattern for putting the coach inside a customer's front door. That means agent-to-agent integration on standard protocols like MCP and A2A, into Copilot, Teams, and customers' own agent platforms, designed once so the fifth customer is configuration.
Write the engineering standards for Forward Deployment. The security bar for customer-facing builds, sandbox and data isolation, customer-data handling, and what "ready to graduate" means. Our code touches HR systems and member data at customers who run us through security review, and the standards are yours to write and hold.
Own the seam with core engineering. Every custom build ends one of three ways: it graduates into the platform with core engineering's acceptance, it converts to paid work, or we kill it. You make the technical case at the monthly graduation board, and you design builds so the ones worth keeping can be accepted cleanly.
Build the shared deployment infrastructure. A real customer sandbox, the tooling that turns a customer's own material into a working context library, and the evaluation harness that tells us whether a configured coach behaves the way that customer needs. Each new account should start from what the last one built.
Qualification
Architecture depth across identityFounders
Required
7+ years building and shipping production software
Several years as technical owner of systems deployed inside large enterprises
Experience leading technical customer deployments with seven or eight figure contracts
Architecture depth across identity, data, and integration
Production experience with LLM and agent systems
Credibility with senior technical audiences
Experience turning one-off customer work into platform capability
Preferred
Senior forward deployed engineers and deployment leads from companies that run this model, who have owned the architecture across several accounts.
Staff and principal engineers or architects from enterprise platforms in identity, HR technology, or data, who know how large customers' systems behave once they're deployed.
Founders, CTOs, and founding engineers who sold into large enterprises and carried the security review themselves.
Applied AI leads who have put agent or LLM systems into production at large companies and answered for how they behaved.
Helping people adapt while AI changes how work gets done is one of the most important problems in the world right now. You'll work on it across entire workforces at many of the world's largest employers, a bigger scale than any one company's internal team gets.
You set the architecture for a function our executive team has named as its expansion lever, on the accounts where the technical answer decides the outcome.
Your work sits on the critical path of decisions worth millions of dollars, and you'll know within a quarter whether it moved them.
You'll build for some of the largest companies in the world, in their environment, with their engineers and architects. Nobody stands between you and the customer. If that's not appealing, this isn't the right role.