> Site index: [llms.txt](https://ortelian.com/llms.txt) with all pages and descriptions. # What is forward deployed engineering? A guide for buyers. Source: https://ortelian.com/forward-deployed-engineering/ Markdown: https://ortelian.com/forward-deployed-engineering.md Forward deployed engineering puts engineers inside your company until AI runs in the work. What FDEs do, how they compare with consultants and when it fits. Forward deployed engineering is a way of delivering software and AI where engineers work inside the customer’s company, with the people who do the work, until the system runs in daily use. A forward deployed engineer (FDE) owns the outcome on site, from the first conversation to the work in production. This guide is for companies deciding whether to bring in forward deployed engineers for AI: what you get, how it compares with consultants and solutions engineers, when it fits and what to ask before you sign. [Talk to us](https://ortelian.com/contact/) [How we work](https://ortelian.com/how-we-work/) Definition ## Engineers who stay until the work runs. A forward deployed engineer is a software engineer who works at the customer’s site. They learn how the work really runs, connect the systems it depends on, build what is missing and stay until the result holds up in daily use. They are measured on whether the work changed. The term comes from Palantir, which embeds engineers with customers to build applications on its software for each customer’s problem. Palantir’s own job description says it [pioneered the position](https://jobs.lever.co/palantir/5168e8fd-fec1-4fea-b7a1-81bdaea65850) and describes FDEs who “embed themselves in the customer’s reality until the problem is theirs”. In 2026 the model spread across AI. [OpenAI launched the OpenAI Deployment Company](https://openai.com/index/openai-launches-the-deployment-company/) in May, [AWS put $1 billion into a forward deployed engineering organization](https://www.aboutamazon.com/news/aws/aws-1-billion-forward-deployed-ai-engineers) in June, and in July [Microsoft launched Microsoft Frontier Company](https://blogs.microsoft.com/blog/2026/07/02/microsoft-frontier-company-ai-engineering-that-amplifies-and-protects-your-intelligence/) and [Anthropic and its partners launched Ode](https://www.ode.com/press/anthropic-blackstone-and-hellman-friedman-introduce-ode-with-anthropic-an-enterprise-ai-services-firm). Each announcement makes the same point: getting access to models is no longer the hard part. Getting AI into the daily work of one specific company is. Our [map of forward deployed AI companies](https://ortelian.com/notes/forward-deployed-ai-companies/) compares who offers what. 1. 01 Map the work Observe it, talk it through with the people who do it and read the records in the systems, including the parts nobody wrote down. 2. 02 Connect the systems Join the data the work depends on, so agents work from the same facts as the team. 3. 03 Build and test Write the agents and workflows, and test them on real examples of the work against a standard agreed up front. 4. 04 Deploy and stay Put it into the tools the team already uses, fix what breaks and keep improving it after launch. Compared ## Forward deployed engineer, consultant, solutions engineer or contractor. Buyers often meet all four in the same quarter. The difference is what each one hands you and what is left for your team to do. For the long version, see [forward deployed engineer vs consultant vs solutions engineer](https://ortelian.com/notes/forward-deployed-engineer-vs-consultant/). Option What you get What stays with you Strategy consultant A recommendation and a roadmap. Building it, running it and finding out whether it works. Solutions engineer A demo and a technical fit check before the sale. Everything after the contract is signed. Contractor or staff augmentation Engineering hours, directed by your team. Deciding what to build and owning the result. Forward deployed engineer Working software in production, built on site with your team. The decisions, and the choice to run it yourselves or keep the team on. Two kinds ## Some forward deployed engineers deploy a product. Others bring a platform and change the work. At a model lab or a software vendor, forward deployed engineers make that vendor’s product work for one customer. They connect the models or the platform to the customer’s data, tools and processes. That is useful if you have already chosen the vendor. At a company that pairs a platform with a forward deployed team, the engineers bring a platform you don’t have to build or host, and use it to change how specific work runs. Ortelian works this way. We are a platform plus a forward deployed team, and the point of both is to get the most impact from AI in your business. We work on site not because our product needs it, but because your company does. The impact comes from changing the work, and that takes time with the people who do it. After launch, your team runs it or we keep running it for you. Either way, we host and operate the [platform](https://ortelian.com/platform/) in the EU. Inside an engagement, your own agents can use it too, through scoped tools we set up for you, and it plugs into the chat tools your team already uses over MCP. See [Ortelian vs ChatGPT Enterprise](https://ortelian.com/ortelian-vs-chatgpt-enterprise/) and [Ortelian vs Microsoft Copilot](https://ortelian.com/ortelian-vs-microsoft-copilot/). How this compares with hiring your own team is on [Ortelian vs building in-house](https://ortelian.com/ortelian-vs-building-in-house/). *Diagram: A code box holding a file tree. Everything the work needs is inside it.* When it fits ## When forward deployed engineering is worth paying for. It fits when the work depends on context no single system holds, when a pilot or a demo worked and nothing changed afterwards, and when nobody inside has the time to sit with the team, build and stay. It also fits when you want one party to own the result instead of splitting the plan, the build and the running across three firms. It is the wrong choice when an off-the-shelf tool already does the job, when you only need a decision on where AI should go, or when nobody inside the company can own the work and make calls. Forward deployed engineers need an internal owner, time with the people who do the work and agreed access to the systems it depends on. How an engagement runs ## Audit, agree the standard, build in two-week cycles, then run it. This is how an Ortelian engagement runs. The full detail is on [how we work](https://ortelian.com/how-we-work/), and the steps are written up in [how to deploy AI agents in an enterprise](https://ortelian.com/notes/how-to-deploy-ai-agents-in-an-enterprise/). 1. 01 An audit on site, about two weeks We map how the work actually runs, find the constraint and redesign the work with your team around what AI can now do. 2. 02 Agree the standard Real examples of your work become evaluations for quality, cost, time and human effort, agreed before anything is built. 3. 03 Build in two-week cycles We build with your team, test on real examples and deploy what meets the standard, in the tools your team already uses. 4. 04 Run it Your team runs it or we keep running it. Either way, we host and operate the platform, and the evaluations keep running. *Diagram: An audit of about two weeks that maps the work and sets the direction, then delivery cycles of two weeks each that build, test and deploy, and keep going. The platform and world model run underneath and carry into every cycle.* Before you sign ## Questions to ask a forward deployed team. - What runs after you leave? The work should run on something you can operate: a platform someone hosts and maintains, with the logic written down where your team can see it. - Who runs it after launch? Ask whether your team can take it over, whether they can keep running it, and what happens to the platform in each case. - How do we know it worked? There should be a standard agreed on real examples before the build starts, and evaluations that keep running after launch. - Which models, and who chooses? A team tied to one model provider will build on that provider. Ask whether models are tested against your work. - Where is our data, and who can act on it? Ask where the platform is hosted, what each agent can access and change, and which actions wait for a person’s approval. Working as an FDE ## Looking for a forward deployed engineering job? This page is for companies buying forward deployed engineering. If you want to work as a forward deployed engineer at Ortelian, see [join us](https://ortelian.com/join-us/). Questions ## Questions buyers ask about forward deployed engineering. **What does a forward deployed engineer do?** A forward deployed engineer works inside the customer’s company to get software or AI into daily use. They map how the work runs, connect the systems it depends on, build and test what is needed, deploy it and stay until it holds up. **Is forward deployed engineering the same as consulting?** No. A consultant usually delivers a recommendation and someone else builds and runs it. A forward deployed engineer builds the system and is measured on whether the work changed. Some consultancies now run their own forward deployed practices, so ask who builds and who stays. **How is a forward deployed engineer different from a solutions engineer?** A solutions engineer helps a customer evaluate and buy a product, mostly before the sale. A forward deployed engineer works after the sale, inside the customer, until the system runs in production. **How long does a forward deployed engagement take?** It depends on the provider and the work. At Ortelian, the audit takes about two weeks on site, then we build in two-week cycles until the first job meets the standard agreed at the start. There is no fixed engagement length. **What happens when the forward deployed engineers leave?** That depends on what they leave behind. At Ortelian, your team runs the work after launch or we keep running it for you. Either way, we host and operate the platform, so nothing depends on code only the engineers understand. **Can our own AI tools use what the engineers build?** With Ortelian, yes, inside an engagement. Your own agents can use the platform through scoped tools we set up for you, and it connects to ChatGPT Enterprise or Microsoft Copilot over MCP. There is no public API or self-serve access. --- ## Let’s see if it fits. Bring one piece of work where AI should be doing more. We’ll tell you whether an audit is worth it. [Talk to us](https://ortelian.com/contact/) [Or email Laurens](mailto:laurens@ortelian.com)