- "Forward deployed engineer" is one title over several different jobs, so the responsibilities decide what the work is.
- The engineer works inside the customer's environment, from discovery through deployment and the first weeks of running it.
- One question splits the postings more than any other: does this engineer write code that reaches production?
- "AI FDE" is not a level or a credential, so read the technologies beside the ownership the posting states.
- Each kind of FDE work is named and counted on the FDE overview, with the postings behind it.
What the work includes
- learn how a customer's process, data and systems work, and turn that into a technical design
- connect APIs, databases, models or internal services
- write and test production code, then deploy and debug it inside the customer's environment
- explain design choices to customer engineers and other stakeholders
- watch how the solution behaves after launch and fix what breaks
- turn repeated customer needs into reusable code or product requests
Some roles add discovery, evaluation or proof-of-concept work before a customer commits. A pre-sales label alone does not tell you how much production work the role carries.
Three shapes behind the same title
| Shape | What the engineer delivers | What to check in the posting |
|---|---|---|
| Builds what the customer runs | A working system the customer operates, with the engineer holding the technical decisions through delivery | Production access, incident duty, who owns the code after launch |
| Integrates and configures it | Data movement, system connections, platform configuration and the application around them | Which APIs and data stores, who signs off the cutover, who fixes a broken integration |
| Proves it works before the sale | A tested evaluation or a small working version that supports a decision | Whether the same person hardens it for production or hands it to a delivery team |
The boundaries overlap. A Databricks FDE posting describes end-to-end data, AI and application solutions with production delivery and customer architecture. A Google Partner FDE posting asks the engineer to co-build agentic workflows, build evaluation and observability systems, and turn field patterns into reusable modules. The employer's own posting is the authority for its requirements.
An AI FDE is the same delivery work on AI systems
Model and prompt evaluation, retrieval and data pipelines, agent workflows, observability, security controls, integration with customer systems. Look for who owns deployment, what reaches production, and who responds when the system fails.
Five questions that settle whether a posting is FDE work
- Who receives the work: external customers, internal teams, or both?
- Does the engineer write production code, configure integrations, build prototypes, or present a product?
- Which project stages are included: evaluation, proof of concept, implementation, operations?
- Does the engineer work inside a customer environment and own technical delivery there?
- What happens to repeated field lessons, and who decides what becomes part of the product?
For the nearby titles, read FDE vs software engineer and FDE vs solutions engineer.