Forward Deployed Engineer

A Forward Deployed Engineer isn't someone who builds new AI models — they take a powerful AI model that already exists and physically go to a client's offices to wire it into that company's data, aging systems, and real-world workflows so it actually runs.

It's a bit like the difference between the person who builds a great washing machine and the installer who has to connect it to each household's cramped, oddly-shaped plumbing so it actually works. However good an AI model is, dropping it straight into a company's old software, quirky data formats, and layered approval processes usually doesn't work on its own. To close that gap, companies station engineers inside client offices rather than at headquarters, having them sit with non-technical staff to solve problems and write code on the spot.

The title comes from a model Palantir has used for a long time, and more recently Microsoft, Meta, Google, OpenAI, and Anthropic have all been hiring for roles under the same name. The role has grown in importance as the ability to actually embed a model inside a business has become as critical as competing on the model's raw performance.

Articles report things like "job postings for Forward Deployed Engineers are up more than 1000%." It's easy to mistake this role for AI researchers who build models, but the key difference is that they take already-completed models and deploy and operate them on-site at client companies, fitted to existing systems.