FDE Comp Explorer
Forward Deployed Engineer compensation, decomposed. Pick an employer tier (frontier lab, Palantir, AI infra, vertical startup, big tech, defence), a level and a market, and see the base/equity/bonus split rather than one headline number — plus the reported band, the confidence label on the underlying source, and the six questions that change what any offer actually means. A transparent multiplier model with its anchors and sources stated, not a scraped table.
An entire deployment, end to end.
Where frontier-lab FDE roles are concentrated. Every headline number you have seen is from here.
Equity-dominant, and that equity is an estimate about a future liquidity event. The bar is the highest in the market: multi-hour take-homes, a defended design review, and a values round that fails more candidates than the coding does.
Anchored on: Perspective 2026 FDE compensation report (~1,200 respondents)
Equity is 62% of this package. That portion is an estimate about a liquidity event that has not happened. Compare offers on the cash line first, then decide separately how much you believe the equity.
Six questions that change what these numbers mean
- Is any of this variable, and on what? Quota or OTE means a pre-sales-shaped role, whatever the title says.
- Units, strike price, latest preferred price, last round date, vesting, refresher policy.
- Travel expectation — in writing. Travel is uncompensated cost.
- How many customers at once? One deep engagement is a different job from six shallow ones.
- Is any part of my review utilisation-based? The clearest early signal of the services trap.
- Who is on-call, and for how long after handoff?
This is a transparent model, not scraped data. Each tier has a published anchor for a senior FDE in a major US market, scaled by a level multiplier and a market multiplier, then shown as a band whose width reflects how good the underlying source is. Frontier-lab and startup figures come from self-reported survey and anecdotal data and are directional only; the Palantir anchor is self-reported levels.fyi data. Verify anything decision-critical against the posting’s own range where pay transparency applies — that is first-party data and it beats everything on this page.
Search “forward deployed engineer salary” and you get numbers that differ by roughly seven times, and both ends are real. One job-board aggregate lands near $156K; a screenshot on your timeline says $1M. They are measuring different tiers of a genuinely bimodal market, and they are also measuring different things — base salary in one case, total compensation in the other. This explorer exists to make that structure visible: pick the tier, the level and the market, and the number decomposes into base, annualised equity and bonus rather than collapsing into a single figure that hides where the money actually comes from.
The model is deliberately transparent. Each of the six tiers carries a published anchor for a senior FDE in a major US market, which is then scaled by a level multiplier and a market multiplier. The Palantir tier is anchored on self-reported levels.fyi data and is the most legible of the six, because Palantir is public and its equity can be valued today. The frontier-lab and startup tiers rest on a practitioner survey and on scattered anecdotal reports, so they are shown with wider bands and a weaker confidence label — the honest position is that this data is directional. Nothing here is scraped live, and the tool says so on the page rather than in a footnote.
The most useful thing the decomposition shows is the equity share. At the top tiers equity is reported at 55 to 70 per cent of total compensation, which means most of the headline number is a forecast about a liquidity event that has not occurred. That is why the tool surfaces a separate cash line and prints the six offer questions alongside the numbers: whether any of the package is variable, the full equity terms, the travel expectation in writing, how many customers you carry, whether utilisation appears in your review, and who is on-call after handoff. Those answers change what a number means far more than the number itself does. For the full treatment of tiers, levels and how to read an offer, the FDE salary handbook carries the sources and confidence labels in detail.
How it works
- Six employer tiers, five levels, four markets.
- Shows base / equity / bonus split, not one headline number.
- Every tier carries its source and a confidence label.
- Copyable summary card with the six offer questions.
Frequently asked questions
Why do published FDE salary figures differ so much?
Because three different things get compared as if they were the same number. Most public figures report base salary while the headline figures report total compensation, and equity is reported at 55 to 70 per cent of a frontier-lab package. Around a third of postings using the FDE title are pre-sales-shaped roles with different pay structures, which drags job-board aggregates down. And the highest-paying roles are geographically concentrated in a few US cities, so a global average blends them with markets paying a fraction as much.
Does tier or level matter more for FDE compensation?
Tier, by a wide margin. A senior FDE at a vertical AI startup can earn less than a mid-level FDE at a frontier lab. Optimising which tier you are in — and the quality of the equity within it — dominates level progression for the first several years of the career, which is why this tool puts tier first.
Where do these numbers come from?
Each tier carries a published anchor for a senior FDE in a major US market, scaled by a level multiplier and a market multiplier. The Palantir anchor is self-reported levels.fyi data; frontier-lab and startup anchors come from a practitioner survey and scattered anecdotal reports and are directional only. Every tier shows its own confidence label and source, and the band width reflects how good that source is. It is a model for orientation, not an offer benchmark.
How should I read the equity portion of an FDE offer?
As an estimate about an event that has not happened. Ask for the number of units, the strike price, the latest preferred price, the last round date, the vesting schedule and cliff, and whether refreshers are standard. When equity is the majority of a package, most of the offer is a forecast about a private company — so compare competing offers on the cash line first, then decide separately how much you believe each equity story.
What should I ask before comparing two FDE offers?
Six things: whether any of the package is variable and on what (quota or OTE means a pre-sales-shaped role whatever the title says); the full equity terms; the travel expectation in writing, since travel is uncompensated cost; how many customers you carry at once; whether any part of your review is utilisation-based, which is the clearest early signal of the services trap; and who is on-call, and for how long after handoff.