The number
60–70%. That's the share of a frontier-lab forward-deployed engineer's total comp that is now equity, not cash — up from 35–45% two years ago, per Perspective AI's 2026 FDE compensation report (1,200 data points; levels.fyi refreshed its blended FDE median to ~$205K on Sep 7). Sit with what that does to the headline numbers everyone quotes. An Anthropic L4 FDE's ~$665K "total comp" is roughly $275K base plus ~$445K of annual equity vest on private stock that liquidates only through occasional secondary tenders. The $785K–$1.2M figures at OpenAI and Anthropic are the same story, more so. So when you read "FDEs clear $600K," read it as: a $250–300K salary, wrapped around a leveraged bet on one company's valuation holding. That's not a knock — it's the single most important thing to model before you sign. A 40% valuation drawdown (the market ran one in 2022) turns that $665K into ~$450K of realizable value. Cash is cash; the rest is conviction.
New roles
Lovable — Forward Deployed Engineering Lead New York / Stockholm (hybrid) · comp not disclosed The successor to Lovable's founding-FDE seat: a player-coach who builds and runs the FDE team in-region while staying hands-on. At a company with 60M+ projects built and enterprise logos like Adidas and Zendesk, this is one of the first jobs defining how the FDE function scales — not whether it exists. View role →
Stability AI — Forward Deployed Engineer Remote (US only; periodic customer travel) · comp not disclosed The company that lit the generative-AI fuse with Stable Diffusion is now staffing FDEs — proof the role has jumped from LLM/enterprise-software land into generative media. You'd embed with creative and enterprise teams to wire image/video/3D models into real production pipelines: VFX, games, ad, e-commerce. A different last mile entirely. View role →
Anthropic — Forward Deployed Engineer (Munich) Munich, Germany (hybrid, 25–50% travel) · comp not disclosed Framed explicitly as "one of our founding FDEs" for the region, and it wants a native German speaker with 8+ years customer-facing engineering. The DACH enterprise market — finance, pharma, industrials — is where European AI budgets actually sit, and planting a local-language FDE there is a sharper signal of enterprise intent than another SF req. View role →
Bunkerhill Health — Forward Deployed Software Engineer San Francisco, CA (on-site, SoMa) · $180K–$285K Top-of-band comp for a healthcare embed. Bunkerhill wants an FDSE to own customer onboarding and on-site clinical deployments — the FDE model pushed into hospitals, where the last mile is HIPAA, legacy EHRs, and clinicians who don't care about your model card. Fully in-person, and the $180K–$285K band is real. View role →
VectorShift — Forward Deployed Engineer New York, NY (on-site) · $150K–$200K + 0.10–0.30% equity The "technical tip of the spear" framing in its purest form: a YC agent-workflow startup wants a full-stack FDE who travels to client sites and writes the code no one's written yet, with finance/PE exposure preferred. A clean, disclosed band plus real early-stage equity — the applied-AI-startup tier from this issue's comp math, made concrete. View role →
Preql AI — Forward Deployed Engineer, Finance Solutions New York, NY (hybrid) · comp not disclosed A women-led finance-data platform (founders ex-WeWork and WaPo data leaders) hiring an FDE to sit inside finance teams — reconciliation and clean reporting without ripping out the ERP. That "embed and automate around the system of record" wedge is one of the most durable FDE beachheads in enterprise. View role →
Tocaro Blue — Forward Deployed Engineer Birmingham, AL (on-site) · $105K–$132K Put this next to the number up top. An investor-backed maritime AI-perception company wants an on-site FDE in Birmingham, Alabama — taking lab models out onto actual vessels. At $105K–$132K it's a fraction of a frontier-lab package, and that's the point: "FDE" now spans a Deep-South $118K maritime seat and a $665K SF agent-builder under one title. Location and vertical, not the words on the badge, set the pay. View role →
One insight: Europe is building its own FDE market
Most of the FDE gold rush has been a US story — SF, NY, DC. This week the board tilted the other way, and the pattern is worth naming: the forward-deployed function is being stood up natively in Europe, not exported to it.
Three of this week's roles make the case. Anthropic's Munich req isn't a remote-friendly afterthought — it wants a native German speaker with 8+ years of customer-facing engineering, and calls the hire "one of our founding FDEs" for the region. That phrasing matters: founding, not fifth. DeepJudge, a Zurich legal-AI scale-up founded by ex-Google search engineers, has gone further and stood up an entire Deployment Engineering org — FDE, pre-sales, production, SRE, solution architect — to roll enterprise search out to firms like Freshfields and Holland & Knight. And Lovable is hiring an FDE Lead in Stockholm, a player-coach to build the team on the ground. Add Mistral's standing FDE reqs in Paris and London, and you have four distinct European employers hiring not "an FDE who happens to be in Europe" but the local nucleus of a field-engineering org.
The logic is the same one that made FDEs a US phenomenon, applied to a different map. DACH and Nordic enterprises — banks, pharma, industrials — sit on the continent's largest AI budgets, and they buy from people who speak their language, sit in their time zone, and can be on-site by lunch. A US-based FDE can't do the embedded part of the job from six time zones away. So the labs and scale-ups are doing locally what Palantir taught them to do at all: put the engineer next to the customer. For candidates, that's the opening. The European FDE market is early enough that "founding" is still in the job titles — and founding seats are where the equity math (see above) and the career math both pay out most.
data: role details per current listings on fde.directory (Anthropic Munich, DeepJudge Zurich, Lovable Stockholm, Mistral Paris/London), verified live this week.
One ask
Which European company should we break down next — a lab, a scale-up, or an enterprise buyer building its own FDE bench? Reply with a name; the best suggestion becomes a company deep-dive.