Key takeaways
- A Forward Deployed Engineer (FDE) is a software engineer who works inside a customer's business, not behind a vendor's wall.
- FDEs combine problem discovery, hands-on engineering and production deployment, and they are measured on outcomes and adoption.
- They are most valuable when the problem is real but poorly defined: for example, automating a business process with AI.
A Forward Deployed Engineer (FDE) is a software engineer who is deployed directly into a customer's organisation to understand its problems first-hand and build working solutions on the customer's own data and systems. Instead of receiving a requirements document and building in isolation, an FDE sits with the people who do the work, turns what they learn into software, and stays until that software is running in production and being used.
Where the role comes from
The term was popularised by Palantir, which built its business around engineers who worked on-site with customers rather than handing over off-the-shelf software. More recently, AI companies, including OpenAI and Anthropic, have built FDE teams for the same reason: large language models and AI agents only create value once they are connected to a company's real workflows, data and systems, and that last mile needs engineers who understand the business.
What a Forward Deployed Engineer does
An FDE's work typically spans the whole journey of a solution:
- Discovery. Shadowing teams, mapping how work actually flows and finding the steps that are slow, manual or error-prone.
- Scoping and prioritisation. Deciding, together with stakeholders, which problems are worth solving first, based on impact, feasibility and data readiness.
- Building. Writing the software: AI agents, LLM and RAG pipelines, machine learning models, integrations and user interfaces.
- Deployment. Shipping to production on the customer's infrastructure with security, monitoring and human-in-the-loop checks.
- Adoption and iteration. Working with users, fixing what gets in their way and measuring results against a baseline.
- Handover. Documenting the system and training the customer's team so they can own and extend it.
FDE vs consultant vs outsourced developer
| Forward Deployed Engineer | Consultant | Outsourced developer | |
|---|---|---|---|
| Where they work | Inside your team | Alongside leadership | At the vendor |
| Starting point | The business problem | The business problem | A specification |
| Writes production code | Yes | Rarely | Yes |
| Deploys and supports adoption | Yes | No | Sometimes |
| Measured on | Outcomes and usage | Recommendations | Delivering the spec |
Skills that make a good FDE
- Strong full-stack and data engineering, so they can build end to end without waiting on other teams.
- Practical AI experience: knowing when an LLM, a classic ML model or a simple rule is the right tool.
- Product sense and communication: translating between business users and technology, and saying no to the wrong problems.
- Ownership: caring whether the solution is used, not just whether it was delivered.
When does your business need an FDE?
Bringing in an FDE makes the most sense when:
- You know AI could help, but you are not sure which processes to start with.
- Previous pilots worked in a demo but never made it into daily use.
- The work depends on your internal systems and data, so an off-the-shelf tool does not fit.
- Your own engineering team is busy and you need extra capacity that can work independently.
If you already have a clear specification and a stable process, a traditional project team may be enough. FDEs shine where the problem has to be discovered as it is solved.
How to measure an FDE engagement
Agree on a baseline before the work starts, then track:
- Time per task or per transaction, before and after.
- Share of cases handled automatically versus by people.
- Accuracy or error rate of the automated output.
- How many people actively use the solution each week.
- Whether your team can run and change the system without the FDE.
Frequently asked questions
What does FDE stand for?
FDE stands for Forward Deployed Engineer: a software engineer who works directly inside a customer's business to build and deploy solutions to the customer's real problems.
Is a Forward Deployed Engineer the same as a consultant?
No. A consultant typically advises and recommends. An FDE also writes the code, integrates it with the customer's systems, deploys it to production and stays accountable for whether it is actually used.
How long does an FDE engagement usually last?
It depends on the goal. A focused discovery-and-prototype sprint can take a few weeks, while embedding an FDE to automate several workflows across a business is usually a longer, ongoing engagement.
Do I need AI expertise in-house to work with an FDE?
No. Part of an FDE's job is to bring that expertise, help you choose the right problems and data, and transfer knowledge to your team so you can own the solution afterwards.
Hire Forward Deployed Engineers from Determinant Studios
Determinant Studios provides Forward Deployed Engineers who embed with your team to discover automation opportunities, build AI solutions on your data and run them in production. Read about how our FDE engagements work or our five-step playbook for AI process automation.
Talk to us about an FDE →