What Is a Forward Deployed Engineer?

The short answer

A forward deployed engineer (FDE) is a software engineer who works directly with customers to get a product working in their environment, on their data and problems. FDEs write production code (integrations, data pipelines, configuration, extensions, AI workflows), and bring what they learn in the field back to the product team.

Where the role came from

The term comes from Palantir, which has long employed Forward Deployed Software Engineers (internally "Deltas") alongside the engineers who build its platforms ("Devs"). Palantir's own description of the difference is that Devs focus on a single capability for many customers, while Deltas enable many capabilities for a single customer (Dev versus Delta; see also A day in the life of a Forward Deployed Software Engineer).

Since 2024, many AI companies and enterprise software vendors have hired forward deployed engineers for the same reason: getting a powerful but general product (a model, a data platform) from an impressive demo to a workflow a customer relies on every day takes engineering work inside the customer's world.

How it compares

week = {"customer workshops and calls": 8, "building integrations and pipelines": 16, "debugging in customer environment": 6,
        "demos and status updates": 4, "writing product feedback and docs": 3, "travel": 3}
total = sum(week.values())
for activity, hours in sorted(week.items(), key=lambda kv: -kv[1]):
    print(f"{activity:38} {hours:3} h  {'#' * hours}")
print(f"{'total':38} {total:3} h  ({week['building integrations and pipelines'] / total:.0%} hands-on building)")

building integrations and pipelines     16 h  ################
customer workshops and calls             8 h  ########
debugging in customer environment        6 h  ######
demos and status updates                 4 h  ####
writing product feedback and docs        3 h  ###
travel                                   3 h  ###
total                                   40 h  (40% hands-on building)

What this tutorial covers

Discovery and scoping, integration engineering (APIs, auth, webhooks, data), deploying in customer environments (security reviews, SSO, networks), AI deployments (retrieval, evaluation, guardrails) and turning prototypes into production. Examples are Python you can run in the browser.