A forward deployed engineer (FDE) is an engineer who works inside a customer's operation instead of behind a product team. They sit with the people doing the work, connect the software to the customer's real data and systems, and stay until it works in production. The job is part engineer, part consultant, part translator between what the business needs and what the technology can actually do.
Palantir made the title famous: engineers embedded with government and enterprise customers to build on live data, not demos. AI companies adopted it as AI moved from chat to action. OpenAI, Anthropic, Salesforce, Databricks and Google all now hire forward deployed engineers, because an AI agent that works in a demo still has to be fitted to each customer's workflows, permissions and data.
Governance written in a policy document doesn't reach the system. An FDE puts it where the work happens: the agent's permissions, the review point before a high-impact action, the receipt every run leaves behind. That's the gap between 'we have an AI policy' and 'we can prove what our AI did.'
Jimmy Standaert works as a Fractional CAIO and forward deployed engineer. The Agent1 case study shows the job in practice: a supervisor agent that followed every rule and still stalled, and the operational fix that got it moving. See also What is Forward Deployed Engineering for AI?
Yes. FDEs write production code, but they spend most of their time in the customer's environment instead of on a central product team.
Yes. A fractional FDE embeds for a defined project, gets one workflow working in production, and hands it over.
Watches the work. The fastest way to build the wrong thing is to skip that step.
