Vibe Engines is free, interactive AI engineering & system design — runs in your browser. Pick a learning path →
Two ladders,
one deployment
Palantir did something the rest of the industry copied halfway: it split forward-deployed work into two distinct roles with different ladders, different interviews and different compensation, and then everyone else adopted the title without the split. That is why “FDE” now means two different jobs depending on who is saying it — and why candidates regularly interview for one while preparing for the other.
The split, as Palantir describes it
Palantir’s own framing is “Dev versus Delta”: the product organisation that builds the platform, and the forward-deployed organisation that makes it work at a customer.13 Inside the forward-deployed side there are then two roles, and the distinction is not seniority — it is what you are accountable for.
| FDSE — Forward Deployed Software Engineer | Deployment Strategist | |
|---|---|---|
| Accountable for | The system working: pipelines, integrations, applications on the customer’s data | The deployment succeeding: the problem chosen, the stakeholders, the outcome |
| A normal day | Data models, transforms, ontology, application logic, debugging someone else’s schema | Working sessions with operators, mapping a workflow, deciding what gets built next |
| Ladder | Engineering | Not an engineering ladder |
| Interview | Coding, decomposition, systems — an engineering loop19 | Case-style, structured problem solving, communication |
| Compensation | Engineering bands, self-reported at roughly $171K–$295K total, median near $211K16 self-reported | A separate track with its own bands |
| Failure mode | Building an excellent system for the wrong problem | Choosing the right problem and being unable to judge whether the build is sound |
Because the two failure modes above are real and they defend against each other. An engineer left alone with a customer optimises the thing they can see; a strategist left alone commits to outcomes nobody has costed. Pairing them is a deliberate structure, and it is the single most-copied and least-understood idea in the whole forward-deployed model — most companies hire the engineer and expect them to do both jobs.
What the rest of the industry did with it
Adopted the title, dropped the pairing, kept the expectations of both.
When OpenAI, Anthropic, and a long list of AI companies adopted “forward deployed engineer”, almost none adopted the two-role structure.910 The result is a single title that absorbs both accountabilities — which is exactly why the role is hard to fill and why the bar is often described as staff-engineer engineering plus a temperament most engineers have never been asked to develop.11
Where a company kept a version of the split
- Two people land at each customer, one engineering-weighted and one outcome-weighted.
- There is a named person whose job is choosing the problem, and it is not the person building.
- The engineer is not the one negotiating scope in the steering committee.
- Job descriptions for the two are visibly different documents.
Where one title carries both
- You do discovery, scoping, building, demos, the security review and the QBR.
- “Owns the full arc of a deployment” appears in the posting.9
- Nobody senior is accountable for whether the chosen problem was worth solving.
- This is the common case, and it is why the role burns people who were hired for one half of it.
“Who decides what we build at a customer, and are they the same person who builds it?” A company with a real split answers immediately. A company without one answers “the FDE, with the account team”, which is a truthful description of a job that contains two jobs. Neither answer is disqualifying — but the second one should change what you negotiate and how you plan your week.
Which one are you being interviewed for?
The loops diverge sharply, and candidates routinely prepare for the wrong one. The signal is in the rounds.
Engineering loop with a deployment accent
Incremental coding where each stage invalidates your last design, a decomposition round scored on your clarifying questions, a learning round on unfamiliar documentation, and a systems discussion grounded in messy customer data.1920 Prepare by building and by practising the rounds — the decomposition and learning simulators cover the two most distinctive.
Case-style loop
Structured problem solving on a business situation, stakeholder judgement, prioritisation under ambiguity, and communication under pressure. Coding is not the axis being measured. Prepare the way you would for a consulting case, plus enough technical literacy to know when an engineer is telling you something is hard.
The shared bar
Every version of this role tests whether you can sit with ambiguity in front of someone who is paying, and produce a next step rather than a caveat. That is what the roleplay simulator drills, and it is the axis on which both ladders actually promote.
Treating Deployment Strategist as a junior FDSE, or as a step toward it. It is a different ladder, with its own progression and its own bands. Moving between them is possible and it is a lateral change of career, not a promotion — and the people who are unhappiest in forward-deployed roles are frequently those who joined one ladder believing it led to the other.
Picking, if you have the choice
| Choose the engineering ladder if | Choose the strategy ladder if |
|---|---|
| You want to keep writing code, and losing that would bother you within a year | You are energised by the room more than by the repository |
| Your satisfaction comes from a thing working that did not work before | Your satisfaction comes from an organisation changing how it operates |
| You want optionality back into product engineering later | You are aiming at product, operations or general management |
| You would rather debug a schema than run a steering committee | You would rather run the steering committee than debug the schema |
Both roles carry a real risk of technical atrophy, and it is worse on the strategy ladder. If you take a forward-deployed role, protect the depth deliberately — the market re-prices you against engineers who spent those years in a codebase, and “I was customer-facing” is not a defence a systems interview accepts. Is the role worth it covers this and the other costs properly.
Whichever ladder, the loop is unusual
The picture: Dev, Delta and the one-title copy
Worked example: one deployment week, split two ways
Take one illustrative 50-hour week in week four of a deployment. The work is the same either way. What changes is who does it. Hours are illustrative; the totals were computed.
| Task | Hours | With the split | With one FDE title |
|---|---|---|---|
| Working sessions with operators | 8 | Strategist leads; FDSE sits in for 2 | FDE |
| Scope and priority in the steering committee | 4 | Strategist | FDE |
| 1:1s with the champion and the blocker | 4 | Strategist | FDE |
| Pipeline and transforms on the customer’s schema | 14 | FDSE | FDE |
| Application logic | 10 | FDSE | FDE |
| Technical answers for the security review | 4 | FDSE | FDE |
| Demo to the champion | 3 | FDSE | FDE |
| Numbers for the business review | 3 | Strategist, from FDSE telemetry | FDE |
| Total | 50 | FDSE 33 h, 31 of them building (94%). Strategist 19 h. | One person, 50 h: 31 building (62%), 19 strategist (38%) |
Under the split, the FDSE spends nearly the whole week building and attends only the sessions that shape the build. Under one title, 38% of the week is a second job, and it is the part that decides what the other 62% is spent on. It also arrives in pieces: a steering committee between two pipeline sessions, a 1:1 in the middle of a debugging afternoon. That is the cost the page means by “a job that contains two jobs”. If you take a single-title role, plan your week around it from day one instead of discovering it in month two.
Cost and trade-offs: split ladders or one title
Here Case is the structure or the choice, Time is what it costs in people or years, Space is what you end up carrying, and Why is the catch.
| Case | Time | Space | Why |
|---|---|---|---|
| FDSE + Strategist pair | Two people per customer | Two ladders, two loops, two comp tracks | Each failure mode checks the other, at the cost of a second hire per deployment. |
| One FDE title | One person per customer | One hire who has to clear both bars | Hard to fill, and it burns people hired for one half of it. |
| Choosing the FDSE ladder | Years of building | Engineering depth; a route back to product engineering | The risk is an excellent system for the wrong problem. |
| Choosing the Strategist ladder | Years in the room | A route toward product, operations or management | Technical atrophy is a bigger risk here. |
| Switching ladders | A lateral move | You leave one progression and start another | Plan it as a change of career, not a promotion. |
Companies that copied the title skipped the first row because it doubles headcount per customer. The saving is real on paper. The cost of skipping it lands on the person in the second row, who now has to be good at two jobs in one week, so ask which row you are joining before you sign.
Common mistakes
- Preparing for the title, not the rounds
Read the loop the recruiter describes. Incremental coding plus a decomposition round is the FDSE shape. A business case and stakeholder questions are the Strategist shape. Ask for the round list if it is not offered.
- Treating the Strategist role as junior FDSE
It is a separate ladder with its own bands. Joining it as a way in to engineering is the unhappy path the page warns about.
- Assuming “FDE” means Palantir’s FDSE
At most companies it means both columns of the table. Expect a week like the right-hand column of the worked example, and negotiate scope and support with that in mind.
- Committing to scope alone in the room
With one title, the builder is also the person asked “can it also…” in the steering committee. Say what it would cost, then decide in writing afterwards. A commitment nobody costed is the Strategist’s failure mode, now yours.
- Letting the other half go untrained
An FDSE who never practises choosing the problem builds the wrong thing well. A Strategist who loses technical literacy cannot tell when an engineer’s “this is hard” is real. Both ladders promote on the shared bar, so train both halves.
Which would you pick?
Three situations. Pick the side you'd actually build — the explanation follows.
Palantir splits the forward-deployed side into FDSE and Deployment Strategist. What is the distinction?
The page states plainly that this is not seniority. An FDSE owns pipelines, integrations and application logic on the customer's data; a strategist owns the problem chosen, the stakeholders and the outcome. Both are demanding, and the second is not the promotion track for the first.
Your loop has incremental coding rounds where each stage invalidates your last design, plus a decomposition round.
The page warns that candidates routinely prepare for the wrong loop, because the signal is in the rounds rather than the title. Design-invalidating incremental coding is the FDSE shape; a strategist loop leans on workflow mapping and stakeholder judgement instead.
How did the rest of the industry adopt Palantir's framing?
The vocabulary spread without the structure, so one company's FDE covers both Palantir roles while another splits them differently. That is exactly why the page keeps returning to responsibilities over labels — the title has become the least reliable part of the posting.
Quick answers
What is the difference between an FDSE and a Deployment Strategist at Palantir?
They are two different ladders, not two levels of one. A Forward Deployed Software Engineer is accountable for the system working — pipelines, integrations, applications on the customer’s data — and sits on the engineering ladder with an engineering interview loop. A Deployment Strategist is accountable for the deployment succeeding: which problem is chosen, the stakeholders, the outcome. That role is not an engineering ladder, is interviewed case-style, and is compensated on its own track.
What does “Dev versus Delta” mean?
It is Palantir’s framing for the split between the product organisation that builds the platform and the forward-deployed organisation that makes it work at a specific customer. The forward-deployed side is then itself split into the engineering role and the strategy role, which is the structure most other companies adopted the title of without adopting the structure.
Is “FDE” at other companies the same as Palantir’s FDSE?
Roughly, but usually with more scope. Most companies that adopted the title did not adopt the two-role pairing, so a single FDE absorbs both accountabilities — discovery, scoping, building, demos, the security review and the business review. That is a large part of why the role is hard to fill and why the bar is often described as staff-level engineering plus a temperament most engineers have not been asked to develop.
Which role should I interview for?
Match the loop to your strengths. The engineering ladder tests incremental coding where each stage invalidates your last design, a decomposition round scored on your clarifying questions, and a learning round on unfamiliar documentation. The strategy ladder tests structured problem solving on a business situation, stakeholder judgement and communication. Both test whether you can sit with ambiguity in front of a paying customer and produce a next step.
Can you move from Deployment Strategist to FDSE?
Yes, and it is a lateral change of career rather than a promotion. Treating the strategy role as a junior version of the engineering one is the most common career mistake in this family of titles — they have separate progressions and separate compensation bands, and people who join one expecting it to lead to the other are consistently the least satisfied.
Sources
Every number on this page traces to one of these. Where a figure is self-reported or crowd-sourced rather than first-party, it is labelled inline.
Cited on this page
- OpenAI — Forward Deployed Engineer job description — “full arc of a deployment”. openai.com/careers/forward-deployed-engineer-(fde)-nyc-new-york-city/
- Anthropic — Forward Deployed Engineer, Applied AI (JD) — 40/30/30 split; MCP servers, sub-agents and agent skills as deliverables. jobs.menlovc.com/companies/anthropic/jobs/69674588-forward-deployed-engineer-applied-ai
- First Round Review — “So you want to hire a forward deployed engineer” — the staff-engineer bar. review.firstround.com/so-you-want-to-hire-a-forward-deployed-engineer/
- Palantir blog — “Dev versus Delta” — FDSE and Deployment Strategist as distinct forward-deployed roles. blog.palantir.com/
- levels.fyi — Palantir FDSE compensation — www.levels.fyi/companies/palantir/salaries/software-engineer/title/fdse
- Exponent — Palantir Forward Deployed Engineer interview guide — www.tryexponent.com/guides/palantir-forward-deployed-engineer-interview
- Blind — Palantir FDSE interview, decomposition round (candidate account) — anonymous first-hand report; lower confidence than a first-party source. www.teamblind.com/post/advice-wanted-palantir-fdse-interview-decomp-round-1ffoysoz
Explore the topic
See this alongside everything else on the same subject — handbooks, system designs, challenges and tools, in one place.
More Handbooks
- The Agentic AI Interview HandbookTwenty topics every senior AI engineer should be able to reason about live — from eval pipelines to reliability patterns for generative systems.Read →
- The Senior AI Engineer Interview Handbook60 questions across architecture, production incidents, agentic systems, RAG, evals, cost, safety, and leadership — what staff-level AI interviewers actually probe for.Read →
- 50 Angular Interview QuestionsA visual handbook covering components, change detection, RxJS, signals, routing, forms, performance, and testing — what interviewers actually probe for in senior Angular roles.Read →
- 50 Python Interview QuestionsFundamentals to advanced: data structures, OOP, iterators & generators, the GIL, asyncio, memory, testing, and the standard library — a visual walk through everything a Python interview touches.Read →
- 51 LLM Evals Interview QuestionsGolden sets, LLM-as-judge, regression testing, offline vs online evals, RAG evals, agent evals, red-teaming, and observability — demystified for interviews and production.Read →
- The Agent Evaluations HandbookA self-contained handbook on evaluating AI agents — theory, interactive widgets, and practical guidance. Trajectory evals, tool-use scoring, LLM-as-judge, observability, and reliability for PMs, engineers, and founders.Read →