The Rise of the Forward Deployed Engineer

Every company now has access to the same frontier AI models. Almost none of them are getting the returns. If the technology is a commodity and the results are not, the difference has to be somewhere else — and increasingly, it comes down to a single role most enterprises have never hired for: the Forward Deployed Engineer.

The numbers behind the gap are stark. Independent research from the RAND Corporation puts the failure rate of enterprise AI projects above 80% — roughly twice that of conventional IT projects. MIT found that 95% of enterprise generative-AI pilots deliver no measurable impact on the bottom line. Dig into why, and it is almost never the model. It is what MIT calls the "learning gap" — the distance between a capable model in a demo and a working system running inside a real operation.

A role born inside the customer's operation

The Forward Deployed Engineer was invented at Palantir in the mid-2000s. Its earliest customers had data so sensitive and idiosyncratic they couldn't even brief a vendor on what they needed. So Palantir stopped asking and started embedding: it put its best engineers inside the customer's environment to observe, build and ship in real time. It called them "Deltas." For a decade, Palantir had more Forward Deployed Engineers than ordinary software engineers.

The name is borrowed from the military — forward deployed means operating at the point of action, not from a base in the rear. That is the whole idea. An FDE isn't a consultant who recommends and hands off, and isn't a solutions architect who advises the people who build. An FDE writes and ships code directly on the customer's systems, and owns the outcome end-to-end. Palantir describes the job as closer to a startup CTO than a contractor.

Why every AI lab is suddenly hiring them

When OpenAI, Anthropic and Google began rolling their models out to large enterprises, they hit exactly the wall Palantir hit two decades earlier — capable model, no path to production. Their answer was identical: place engineers directly with the customer. OpenAI stood up a Forward Deployed Engineering team in late 2024; Anthropic set out to grow its applied-AI team roughly fivefold in a single year. Between January and September 2025, job postings for the role rose more than 800%. It is now the most in-demand role in enterprise AI, with total compensation at the frontier labs reaching into the mid-six figures.

The results, where FDEs are deployed, aren't marginal. When BBVA worked with OpenAI's forward-deployed model, a ChatGPT pilot grew into an AI-native banking capability now serving 120,000 employees across 25 countries. That is the gap between a pilot that impresses and a capability that scales — a gap of delivery, not of technology.

The missing role in your AI transformation

Boston Consulting Group's rule of thumb is that successful AI is 10% algorithms, 20% technology and data, and 70% people and process. That 70% — redesigning the workflow, reskilling the people, earning trust in the output — is precisely where pilots die, and precisely where a Forward Deployed Engineer lives. The role closes the gap in a way no tool can:

• →They build on your operation, not a slide. What ships is a working capability against your real systems and data — not a recommendation for someone else to implement later.
• →They close the learning gap. Embedded in the work, they find the edge cases and exceptions the demo never surfaces — and build for them.
• →They stay until it works. No fixed end date that ends the momentum — they own the outcome in production and iterate until the P&L moves, then hand the capability to your team.

How PPIC brings FDE talent to ASEAN

There's a catch in the current market: the FDEs the model vendors hire are there to drive adoption of that vendor's platform. Most organisations in Singapore and across ASEAN don't need another platform pushed at them — they need embedded technical talent that starts from their operation and outcome, free to use whatever tools serve it. That is the gap PPIC Technologies fills.

As a CoolRIOTS company and an IBM Gold Partner, we treat the Forward Deployed Engineer as a new class of tech talent — one we source, develop and place, and one we deploy directly into engagements to build the hybrid workforce alongside your people. Our engineers embed and build, deploy governed CoolMOBS Digital Employees into the workflow, and transfer ownership so the combined human-and-digital operation keeps running — and improving — after we leave. It's how AI capability finally turns into business value: not by buying a platform and hoping, but by deploying the person who makes it work.