Become an FDE

Are you a proven and experienced applied AI or Forward Deployed Engineer?

Work at the forefront of Applied AI

Do you want to work at the forefront of applied AI technology, helping some of the world’s most well-known global enterprise and mid-market organisations solve real-world challenges with innovative AI solutions?

At FDECo, you’ll have the opportunity to work alongside talented people, tackle exciting challenges and make a genuine impact in a rapidly evolving industry. If you’re ready to be part of what’s next in AI, we’d love to hear from you.

What a Forward Deployed Engineer actually does

You embed with a customer and you own the outcome. Not the ticket, not the sprint, whether the thing works in production inside their organisation.

In practice the split is roughly what the job postings describe: a majority of your time customer-facing, a substantial share writing deployment-specific code, and the remainder internal. You scope ambiguous problems with people who are not engineers, decide what is worth building, build it, and stay with it until it survives contact with real users.

It suits engineers with high agency and a tolerance for ambiguity. It does not suit people who want a clean specification and a quiet room. Both are legitimate preferences — be honest with yourself about which one you are.

Why this role, and why now

Palantir created the role in 2010. It remained unique to the company until AI labs faced the same problem, capable products that customers couldn’t get into production.

Postings rose more than 800% across the first three quarters of 2025 and kept climbing through 2026, and compensation at the top of the market now competes with senior research roles. The gap is not going to be filled by people who already have the title. There are not enough of them. It will be filled by experienced software engineers who add applied AI depth and customer-facing capability to what they already have — which is the entire premise of this programme.

The entry bar is high, deliberately

We upskill experienced software engineers. We do not train people from scratch, and we are not a route into the industry.

You need a real track record of shipping and owning production software. You do not need existing AI experience — that is what the programme builds. You do need to be the kind of engineer who moves toward the ambiguous problem rather than away from it, because no amount of curriculum fixes that. We turn down a significant proportion of applicants. That is not a marketing line — it is how the certification retains any value when you put it in front of a hiring manager.

What the programme covers

The technical half: agent architecture and orchestration, multi-agent system design, retrieval and context engineering, tool and API integration, evaluation and regression testing, guardrails and safety design, observability, and the cost and latency engineering that decides whether a system is viable in production.

Built with leading AI labs and platform vendors, and revised continuously because the field will not sit still. The half most programmes skip: consulting. Discovery, problem decomposition, scoping, stakeholder management, and knowing when to tell a customer that what they have asked for is the wrong thing. Forward deployed work happens inside someone else’s organisation and this is what separates engineers who are effective there from engineers who are merely competent.

You certify by doing the job, not by attending

Every engineer completes a minimum two-week consulting and technology capstone: ambiguous requirements, a real technical problem, stakeholders to manage, and a working system at the end.

It is assessed by a panel of practising applied AI experts. Not everyone passes. If you do, what you are holding is an assessment of your work by people currently doing this for a living — which is a materially different thing from a completion certificate.

Where certified engineers go

Certified FDEs are deployed to enterprises, systems integrators and AI vendors, on permanent, contract or embedded squad engagements. Some engineers come to us already employed, sponsored by their organisation, and go back to lead applied AI work internally.

We work in partnership with Tenth Revolution Group and TRG has spent two decades placing specialist technology professionals across eleven countries. The route from certification to a live deployment is not something we are inventing from scratch.

No. You need substantial production software engineering experience. Applied AI depth is what the programme builds — the engineering foundation is what it assumes.

Compensation varies widely by market, seniority and employer. At the top of the market — frontier AI labs — it competes with senior research roles. Enterprise and partner roles sit lower but well above equivalent internal engineering positions.

No. Bootcamps take people into engineering. This takes experienced engineers into applied AI deployment work, and the entry bar reflects that.

You do not certify. We will tell you where the gap was. We would rather have that conversation than put you in front of a client unprepared.

Not necessarily. Some engineers are sponsored by their employer and return to lead applied AI work internally. Others certify in order to move. Both routes are supported.

Frequently hybrid, and often travel-heavy. Forward deployed means deployed forward — proximity to the customer is a large part of why the model works.

Questions engineers ask

Apply for a cohort

Tell us what you have shipped. If the fit is right we will take you through the assessment process.