Skills · Forward Deployed Engineering (FDE)
A forward deployed engineer — FDE for short — takes a messy problem inside a real company and builds working software for it. Palantir grew fast on this one role. It needs two skills that rarely show up in the same person: real coding ability, and the judgement to use it inside someone else’s mess. That combination is rare. That’s why FDEs are paid so well.
One-page briefs, either one to forward without a login: Campus track (PDF) ↓ · Enterprise track (PDF) ↓
Why this pays
The scarcity argument
This isn’t about a big salary number. It’s simpler than that. Most training programmes teach the engineering, or they teach the client judgement — never both. Enterprises pay a premium for whoever can do both. That person is an FDE.
Most engineers can build. Few can sit with a client who isn’t sure what they want, figure out what they actually need, and build it inside someone else’s rules. That combination is naturally rare — not an accident.
Palantir grew fast on this one role. Enterprises now try to hire FDEs directly, instead of paying outside contractors. A role this hard to fill costs more.
Every enterprise adopting AI hits the same wall: it works in the demo, and nobody can get it running on their real systems. That gap only closes when an FDE is in the room.
Pick your track
Same spine, different entry
Campus starts with students who can code, but have never built something for a real client. Enterprise starts with engineers who already have. Each course teaches the part that’s missing.
For students who can already code but haven’t built something for a real client yet. Runs inside a university partnership, one cohort per term.
12 weeks, one term · 20 per campus cohort
For engineers who already ship production code but have never had to do it inside a client’s rules. Paid for by your company, run against your own problem or one from the bank.
6 weeks, full-time or near-full-time · 6–10 per enterprise cohort
| Campus track — Foundational | Enterprise track — Advanced | |
|---|---|---|
| Audience | Universities, via a university partner | Enterprises, Startups & Corporates — sponsoring their own engineers |
| Format | Cohort, one evening a week plus project time, alongside a full course load | Intensive cohort, full engagement, on-site or remote with the sponsoring team |
| Duration | 12 weeks, one term | 6 weeks, full-time or near-full-time |
| Cohort size | 20 per campus cohort | 6–10 per enterprise cohort |
| Capstone | A working prototype against a real, anonymised case from the IIDA problem bank — not a toy dataset, and not a simulated client. | A pilot deployed against a live problem — either the sponsoring enterprise’s own, or one from the IIDA problem bank matched for fit. |
| Cost | Paid for by the university under our academic partnership — free for students | Paid for by your company, priced per cohort — like any project you’d hire us for |
How every cohort runs
Every capstone uses a real, anonymised case from the IIDA problem bank — the same cases our sponsored research is judged on.
Instructors are IIDA researchers and engineers who’ve done this kind of work themselves — not full-time teachers reading from a syllabus.
Cohorts stay small enough for real review of real work. Applying does not guarantee a seat, in either track.