OpenAI is building an entire company around it. So is Anthropic. Google, Mistral and Cohere are recruiting at full speed, and the job listings have exploded over the past year. The title is the same everywhere: Forward Deployed Engineer. It is the hottest role in AI right now. The concept itself is anything but new.
Palantir invented the role almost twenty years ago
The term comes from Palantir, which early on faced an unusual problem: its customers in American intelligence could not explain what they needed. The requirements were classified, and so were the processes. Palantir’s solution was to move the engineers out to the customer. They sat in the customer’s environment, saw the problems first hand and built the solution where it would be used, with responsibility all the way from the first needs analysis to operations. The model became so central that Palantir long had more Forward Deployed Engineers than regular software engineers.
The rest of the industry mostly treated it as a curiosity: expensive, hard to scale and tightly coupled to Palantir’s own platform. The consulting world carried on as usual, with pre-studies, reports and handovers in PowerPoint.
Then came AI, and suddenly the model fit perfectly
Generative AI has a well-known production problem. The models ace the demo, but faced with real data, undocumented workflows and existing systems, most initiatives stall. An MIT study from 2025 reviewed hundreds of AI pilots at large enterprises and found that roughly 95 percent produced no measurable impact on the bottom line. The researchers’ conclusion was telling: the fault was not in the models, but in the integration.
That is exactly the problem a Forward Deployed Engineer solves. An engineer who sits in the customer’s team sees how the work actually gets done, finds the processes worth automating, and can build, evaluate and deploy in the environment where the solution will live. No requirements specification in the world beats sitting next to the person doing the job today.
The AI giants have drawn the same conclusion. In May 2026, OpenAI launched The Deployment Company, a company with over four billion dollars in capital and a single mission: to take the models all the way into enterprise operations. Just days earlier, Anthropic presented a corresponding initiative together with Blackstone and Hellman & Friedman, among others. The model companies themselves have realised that models do not deploy themselves.
The crucial difference: platform or ownership
There is a nuance here that is easy to miss. At Palantir, the role existed to make customers succeed with Palantir’s own platform. The engineers were brilliant, but the end goal was always for the business to land in the vendor’s product.
What has happened now is that the way of working has been decoupled from the platform. When AI needs to be adapted and integrated into a real business, the embedded model is superior, regardless of whose models are used. And then the next question becomes the decisive one: who owns what has been built when the engineers go home?
How we work
For us at TokenTek, this is not a trend to jump on. It is how we have worked from the start: our engineers work at the customer, in the customer’s teams and infrastructure, and train the customer’s people as the solution takes shape. The difference from the platform model lies at the end of the journey. When we hand over, the customer owns everything: models, code, documentation and the way of working.
The industry has finally put a name on how we work. We take that as a compliment.
Want to see what it means in practice? Get in touch and we will show you.