You might have noticed a new job title turning up in your feed over the last year or so: the forward deployed engineer, or FDE. If you have sat through a vendor meeting about AI recently, there is a fair chance one has already been offered to you, usually as part of the deal.
The idea is simple enough. Instead of selling you software and leaving you to work out how to use it, the vendor sends an engineer to sit with your people, often in your offices, to configure the tools, connect them to your systems and get something working against a real business problem. Palantir made the model famous, and the big AI names, OpenAI, Anthropic etc., now all have openings for the role. The growth is the kind that sounds almost made up: Lightcast data reported by Fortune puts postings for forward deployed engineers at more than eleven times last year’s level, up over 1,000% between January and August against the same months of 2025, while the wider tech job market grew by about 13%.1
What caught my eye was a line from one of Palantir’s own job listings, quoted in the same piece: the FDE’s responsibilities “look similar to those of a startup CTO”. As someone who has held the CTO seat a few times, my first reaction was mild indignation; my second, after reading what the job actually involves, was that they have a point.
Strip away the title and much of the FDE’s day is familiar. You start with the business problem and come to the technology second. You find out how the work really gets done, which is usually not quite how the process document says it gets done. You translate between the people who run the business and the people who build the systems, and you make the trade-offs about what to build now, what to buy and what to leave alone. Above all you get something into production rather than into a slide deck, which is where a lot of AI work goes to die. That is most of what I would describe as good CTO work, and it is close to how I run a diagnostic: talk to the people, sit in on the work, look at the data.
Where the two part company is ownership. The FDE works for the vendor; their job is to make one product succeed at your site, and the good ones are very good at it. The CTO’s job is the whole portfolio: every vendor, the team, the budget, the risk, the roadmap and the occasional unpopular decision to stop something. An FDE will seldom tell you that their own product is the wrong answer, and it would be odd to expect them to. The bigger issue comes when they leave. Whatever they built inside your business, the integrations, the prompts etc., has to belong to someone on your side, because unowned capability creates debt for someone later.
So will the roles merge? I suspect they will drift closer together without becoming one job. The FDE boom is the industry quietly admitting that AI does not deploy itself; the hard part has usually been the organisation around the tool, and vendors have worked out that someone has to be in the room for it to stick. That pulls CTOs towards the FDE end of the job, closer to the work and more hands on with how AI lands in each team, and I expect some businesses to build their own forward deployed capability reporting to the CTO, alongside the vendor’s. In smaller businesses the line is already blurry; a fractional CTO inside an SME spends a good part of the week doing something that looks a lot like forward deployment, just on behalf of the business rather than the vendor.
If you have vendor FDEs on site, or are about to, the practical step is small and worth taking early: name the person on your side who will own what they build, and put them in the room from the first week. The FDE will make the product work; whether it keeps working after they move on to their next customer is down to you, and that is much cheaper to settle before they go.
As for whether the FDE is the new CTO, I would put it the other way round; the best CTOs I have worked with were already a little forward deployed.