By FDE Interview Prep, a SourceStrongAI company

In this post11 sections
  1. What Palantir’s Build to Apply postings ask for
  2. Who it is for: the Deployment Strategist, not the FDSE
  3. Pick a problem a customer would pay to fix
  4. Scope one working workflow
  5. Using AI while you build
  6. Script the video
  7. What the video should prove about you
  8. Practice the part the video can’t fake
  9. Questions people ask
  10. Keep reading
  11. More from the blog

You found a Palantir posting that wants to see something you built before it talks to you, and it tells you to use AI while you build it. Palantir’s Build to Apply postings, for its US Government roles, ask you to build a project on Foundry and AIP and submit a demo video, and they say showing your work up front can mean expedited consideration. Source 1Deployment Strategist, Build to Apply - US Government (Washington, D.C.) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job posting The video must be under 5 minutes, sent as an unlisted YouTube link, within 7-10 days of when you begin building. Source 2Deployment Strategist, Build to Apply - US GovernmentPublisherPalantir Technologies (Lever)Source typecompany job posting The judgment it tests, turning a vague problem into one working workflow, is the one the FDE interview guide is built around. Below: eligibility, setup, the problem filter, a build plan for the window and a timed video script.

What Palantir’s Build to Apply postings ask for

As of September 2026, Palantir lists six “Deployment Strategist, Build to Apply - US Government” postings, first posted in February 2026, in Washington, D.C., New York, San Diego, Colorado Springs, Fayetteville and Honolulu. Source 1Deployment Strategist, Build to Apply - US Government (Washington, D.C.) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 2Deployment Strategist, Build to Apply - US GovernmentPublisherPalantir Technologies (Lever)Source typecompany job postingSource 3Deployment Strategist, Build to Apply - US Government (New York, NY) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 4Deployment Strategist, Build to Apply - US Government (Colorado Springs, CO) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 5Deployment Strategist, Build to Apply - US Government (Fayetteville, NC) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 6Deployment Strategist, Build to Apply - US Government (Honolulu, HI) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job posting Here is what they ask, in their own terms:

The postings also say what the video should explain. They want to understand why you selected the problem, how you approached the use case with their products, and the solution you developed. Source 1Deployment Strategist, Build to Apply - US Government (Washington, D.C.) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 2Deployment Strategist, Build to Apply - US GovernmentPublisherPalantir Technologies (Lever)Source typecompany job posting Build your script around them.

“Potential” is the posting’s word: the build is a way to be seen sooner, not a promise of an interview. We found nothing on who watches the video or what happens after review.

Getting set up

The postings answer the first practical question, which is how to get Foundry at all: Source 1Deployment Strategist, Build to Apply - US Government (Washington, D.C.) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job posting

  • Create a DevTier account with AIP, and build the project in it.
  • Take the crash course at learn.palantir.com if Foundry is new to you.
  • Ask questions in the Palantir Developer Community.
  • Bring data: publicly available data, or notional data you create yourself.

The posting counts the window from when you begin building, and doesn’t define that. So our advice is to do the thinking first: pick the user, find the data and write the one-sentence problem before you open your DevTier account.

Who it is for: the Deployment Strategist, not the FDSE

Check eligibility before you spend a week building. As of September 2026, four of the six postings (Fayetteville, San Diego, Honolulu and Colorado Springs) require an active US security clearance at or above the Top Secret level. Source 1Deployment Strategist, Build to Apply - US Government (Washington, D.C.) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 2Deployment Strategist, Build to Apply - US GovernmentPublisherPalantir Technologies (Lever)Source typecompany job postingSource 3Deployment Strategist, Build to Apply - US Government (New York, NY) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 4Deployment Strategist, Build to Apply - US Government (Colorado Springs, CO) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 5Deployment Strategist, Build to Apply - US Government (Fayetteville, NC) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 6Deployment Strategist, Build to Apply - US Government (Honolulu, HI) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job posting The Washington, D.C. and New York postings accept an active clearance or eligibility and willingness to obtain one. Source 1Deployment Strategist, Build to Apply - US Government (Washington, D.C.) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 2Deployment Strategist, Build to Apply - US GovernmentPublisherPalantir Technologies (Lever)Source typecompany job postingSource 3Deployment Strategist, Build to Apply - US Government (New York, NY) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 4Deployment Strategist, Build to Apply - US Government (Colorado Springs, CO) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 5Deployment Strategist, Build to Apply - US Government (Fayetteville, NC) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 6Deployment Strategist, Build to Apply - US Government (Honolulu, HI) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job posting Five of them require the ability to travel 25 to 75%, and the Colorado Springs posting asks for 4 to 5 days a week on customer sites in that area instead. Source 1Deployment Strategist, Build to Apply - US Government (Washington, D.C.) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 4Deployment Strategist, Build to Apply - US Government (Colorado Springs, CO) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job posting As of September 2026, all six estimate the salary at $110,000 to $170,000 a year, an estimate that leaves out sign-on bonus, benefits and long-term incentives. Source 1Deployment Strategist, Build to Apply - US Government (Washington, D.C.) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 4Deployment Strategist, Build to Apply - US Government (Colorado Springs, CO) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job posting

Build to Apply is not the application. On Palantir’s Lever board the FDSE posting sits under the team “Delta” and the Deployment Strategist posting under “Echo”. Source 7Forward Deployed Software EngineerPublisherPalantir Technologies (Lever)Source typecompany job postingSource 8Deployment StrategistPublisherPalantir Technologies (Lever)Source typecompany job postingSource 9Software Engineer - Apollo PlatformPublisherPalantir Technologies (Lever)Source typecompany job posting For the difference in the day job, read Deployment Strategist vs forward deployed engineer, and for the interview questions our sources tie to Palantir, see Palantir FDSE interview questions.

The strategist posting explains why a demo is a fair test. It says strategists go onsite with customer analysts, work with Forward Deployed Engineers to integrate data into a pipeline, and bring experience with programming, scripting or statistical packages. Source 8Deployment StrategistPublisherPalantir Technologies (Lever)Source typecompany job posting Palantir’s Deployment Strategist postings also list building and delivering demos to new and existing customers. Source 8Deployment StrategistPublisherPalantir Technologies (Lever)Source typecompany job posting The Build to Apply video is a sample of that work: you scoped something, built it and showed it to a customer.

Building a demo can also be an interview stage. One candidate on Blind reported, in March 2026, that with no prescreen their first Palantir Deployment Strategist interview stage was building a demo app with Foundry and AIP, listed on their candidate portal as “Mid-Stage Interviews”. Source 10Palantir Deployment StrategistPublisherBlind (teamblind.com)Source typecandidate report on Blind The poster did not say whether this was the Build to Apply route. Either way, be ready to walk through your build live if you are asked.

If you are applying for the FDSE role instead, what Palantir FDSE candidates report covers that loop.

Pick a problem a customer would pay to fix

Palantir’s own annual report describes AIP bootcamps that deliver real workflows on actual customer data in days, which is the same thing the postings ask you to show. Source 11Palantir Technologies Form 10-K for fiscal year 2025PublisherPalantir Technologies (SEC EDGAR)Source typecompany SEC filing A generic dashboard over a famous dataset shows much less.

Here is the filter we use to choose a problem. It is our method, not something Palantir publishes. Keep an idea only if it passes all four:

  1. A named user. A job title, not “stakeholders”: a fire inspection scheduler, a grants officer, a warehouse planner.
  2. A recurring decision. Something that person decides every day or week, such as which sites to inspect, which claims to review first, or which trucks to send.
  3. Data you can get legally. Public data is safest: a city’s open data portal, federal spending data from USAspending.gov, or disaster declarations from OpenFEMA. Or use notional data you generate yourself, which the postings explicitly invite. Source 1Deployment Strategist, Build to Apply - US Government (Washington, D.C.) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job posting Never use data from an employer or client.
  4. A place for AIP to do real work. Something a language model can read, draft or recommend inside the decision, with a person checking it. If AIP only decorates the app, pick again.

The postings are for US Government roles, so a public-sector user fits naturally, though nothing we found requires one. Here is the filter applied to three first ideas:

First ideaProblemRewrite
A dashboard of all federal spendingNo user, no decisionA grants officer ranks which awards to review first
An AI chatbot over city documentsThe decision is unclearA permit reviewer gets a drafted deficiency letter to approve
Fire inspections ranked by riskPassesKeep it, name the scheduler, add an AIP briefing

The rewrites are what a strategist does in a first customer meeting: turn “we want AI” into one person, one decision and one measure. Our lesson on stakeholders and success metrics covers choosing the measure the customer already trusts, and the free question on building one feature from an app’s event data is the same move in miniature.

Scope one working workflow

Take the fire inspection idea. The scheduler for a city fire marshal’s office picks, each week, which businesses to inspect. Today that means a spreadsheet and memory. Your build gives them a ranked list with reasons, an AI-drafted briefing for each visit, and a way to schedule from it.

Scope it as a : every layer, thin, joined end to end. In Foundry terms, that is roughly:

  • Data in. Load the city’s inspection and violation history and its business license list, and clean them in a pipeline. Keep a note of every cleaning rule.
  • Objects the user thinks in. Model a Business, an Inspection and a Violation as object types in the Ontology, with the links between them.
  • A screen they act in. One app view: the ranked list, and a detail panel showing why each business is on it.
  • One action that writes back. “Schedule inspection” records the decision, so the tool changes what happens next instead of only displaying data.
  • The AIP step, at the center. An AIP Logic function reads the Business, its Inspections and its Violations from the and drafts two things: a short briefing for the inspector and a suggested focus for the visit (“check the kitchen hood suppression; it failed twice”). Its edits are staged for the scheduler to review, not applied blindly. Palantir’s AIP Logic overview says Logic edits can be automatically applied or staged for human review.

That is the whole build. No login pages, no second use case, no model training. The walking skeletons lesson explains why thin and complete beats deep and partial, and what you may fake. In a demo, you may fake the volume of data and the number of users. You may not fake the workflow or the AI output: if the briefing is wrong on some inputs, show one and say why.

A plan for the window, which is ours to adapt, not Palantir’s:

  • Before you start. Pick the user, find or generate the data, write the one-sentence problem.
  • Day 1 and Day 2. Data in, cleaned, and modeled as objects.
  • Day 3. The app view and the write-back action.
  • Day 4 and Day 5. The AIP Logic function: what it reads, what it drafts, the prompt, and the staged review. Test it on ten businesses and note where it goes wrong.
  • Day 6. Break it on purpose: missing fields, a business with no history, a briefing that invents a violation.
  • Day 7. Script, record, watch it back, submit. Anything after that is spare time, not scope.

The AIP step gets two days because the postings say it is assessed. Source 1Deployment Strategist, Build to Apply - US Government (Washington, D.C.) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 2Deployment Strategist, Build to Apply - US GovernmentPublisherPalantir Technologies (Lever)Source typecompany job posting

Using AI while you build

The postings are explicit: LLMs, AI and other technical tools are permitted and encouraged while you build. Source 2Deployment Strategist, Build to Apply - US GovernmentPublisherPalantir Technologies (Lever)Source typecompany job posting Use them where they save the most time and check everything they produce.

Good uses. Drafting pipeline transforms and cleaning code. Explaining an unfamiliar Foundry feature. Generating notional records and ugly test cases: a business with no address, an inspection dated in the future. Tightening your script after you have written it.

Where to stay in charge. Picking the problem, choosing the ranking rule, and deciding what the AIP step may do on its own. Those are the judgments the video exists to show, so they should be yours.

Say it in the video. One sentence is enough: “I used an assistant to draft the cleaning code and the test records, and I checked each rule against the raw data by hand.” That turns AI use from a question into a sign of good process.

Keep two kinds of AI apart when you talk. The coding assistant helped you build; AIP is part of the product. Only the second is the capability the postings grade.

The encouragement covers the build. For any later live round, ask your recruiter before you assume.

Script the video

The limit is under 5 minutes, so aim to finish with time to spare. Source 2Deployment Strategist, Build to Apply - US GovernmentPublisherPalantir Technologies (Lever)Source typecompany job posting Film it as a customer demo, not a tour of what you clicked. The times below are our suggestion.

0:00 The user, the decision, and why. Open on the app, not on Foundry’s home screen. “This is for the scheduler in a city fire marshal’s office. Every week they choose which businesses to inspect, and today that’s a spreadsheet and memory. I picked this because the inspection data is public and the decision repeats every week, so a better tool pays off every Monday.”

0:30 The data, briefly. “It runs on the city’s public inspection history and business licenses, cleaned in a pipeline and modeled in the Ontology as businesses, inspections and violations.” Show the object graph for a few seconds, then move on.

1:00 The live run, with AIP named. Open the ranked list and pick the top business. Then say it: “This is the AIP Logic function. It reads this business’s violation history from the Ontology and drafts a briefing and a suggested focus for the inspector. Nothing is saved until the scheduler approves it.” Edit one line, approve it, and schedule the inspection. Show the record update.

2:45 One decision. Name the choice a reviewer would question, the alternative and why. “I ranked by time since last inspection and the severity of past violations, not by a model score, because a scheduler has to explain the list to a business owner who asks why they were picked. I used AIP for the briefing, where a wrong word gets caught in review.”

3:30 Where it breaks, and what’s next. Show one failure. “For a business with no history, the briefing has nothing to go on, so it says so instead of guessing. With the office’s own data I’d add license age and complaint calls.” End on the next step you would take with this customer.

4:15 Stop. That leaves margin under the limit for a slow page load.

For recording mistakes and exact words for trade-offs, the take-home walkthrough video post has a shot-by-shot script. Three points apply to this video in particular:

  • Watch the link signed out. Open the unlisted YouTube link in a private window before you submit.
  • Zoom in. Foundry screens are dense. Zoom the browser until a phone viewer could read the list.
  • Hide what is not yours. Close tabs, notifications and anything not public or notional.

What the video should prove about you

Map the script onto the posting’s own words. It wants to understand why you selected the problem, how you approached the use case with its products, and the solution you developed, and it assesses how you integrate AIP into the demo. Source 1Deployment Strategist, Build to Apply - US Government (Washington, D.C.) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 2Deployment Strategist, Build to Apply - US GovernmentPublisherPalantir Technologies (Lever)Source typecompany job posting

What the posting asks to seeWhere your video shows it
Why you selected the problem0:00, in one sentence
How you approached it with Foundry and AIP0:30 to 2:45
The solution you developedThe live run from 1:00
AIP integrated into the workflowThe Logic function, named at 1:00

Before you upload

  • You have checked the clearance line on your location’s posting
  • The first sentence names a user, a decision and why you picked it
  • You say “AIP” aloud, and the AIP step reads from the Ontology
  • The AI output has a human check, and one miss is shown
  • It runs under the posting’s limit, and the link opens signed out

Practice the part the video can’t fake

The build takes days. The judgment behind it, turning a vague request into one user and one decision, is what every strategist and round comes back to, and you can practice it today. Work through improving a city’s bus reliability as a warm-up, and study the worked case on a city’s building permits for the full method on a public-sector problem.

Then do it against someone who answers back. The free practice case puts you in front of an AI customer with a vague problem and hidden facts; it needs a sign-in. Pro opens the full curriculum and question bank, and starts with a 7-day free trial. See Pro when you want the rest.

GlossaryDeployment strategistA customer-facing role that works out the customer’s questions and scope beside FDEs; at some employers the title means a product-manager or quota-carrying role instead.More on Deployment strategistGlossaryForward deployed software engineerPalantir’s title for its FDE role, called Delta internally; OpenAI and EY also post FDSE titles, each with its own duties.More on Forward deployed software engineerGlossaryWalking skeletonThe thinnest end-to-end version of a system that performs one small real function across its main components, built first and then extended.More on Walking skeletonGlossaryOntologyA shared model of a customer’s business objects and their relationships, mapped onto the underlying data so people and applications use the same terms.More on OntologyGlossaryForward deployed engineerA software engineer who builds and ships production systems inside a customer’s problem and environment, accountable to that customer’s outcome.More on Forward deployed engineer

Questions people ask

What is Palantir Build to Apply?

It is an application route for Palantir’s US Government Deployment Strategist roles. The postings ask applicants to build a Foundry and AIP project and submit a demo video, and say showcasing the work up front can mean expedited consideration.Source 1Deployment Strategist, Build to Apply - US Government (Washington, D.C.) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 2Deployment Strategist, Build to Apply - US GovernmentPublisherPalantir Technologies (Lever)Source typecompany job posting

Do I need a security clearance for Palantir Build to Apply?

It depends on the location. As of September 2026, the Fayetteville, San Diego, Honolulu and Colorado Springs postings require an active US security clearance at or above the Top Secret level. The Washington, D.C. and New York postings accept an active clearance or eligibility and willingness to obtain one.Source 1Deployment Strategist, Build to Apply - US Government (Washington, D.C.) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 2Deployment Strategist, Build to Apply - US GovernmentPublisherPalantir Technologies (Lever)Source typecompany job postingSource 3Deployment Strategist, Build to Apply - US Government (New York, NY) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 4Deployment Strategist, Build to Apply - US Government (Colorado Springs, CO) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 5Deployment Strategist, Build to Apply - US Government (Fayetteville, NC) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 6Deployment Strategist, Build to Apply - US Government (Honolulu, HI) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job posting

How do I get access to Foundry for Build to Apply?

The postings tell you to create a DevTier account with AIP and use it for your project. They point to learn.palantir.com for a crash course and to the Palantir Developer Community for questions, and they welcome public data or notional data you create yourself.Source 1Deployment Strategist, Build to Apply - US Government (Washington, D.C.) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job posting

What does Palantir assess in the Build to Apply video?

The postings say your workflow must use AIP as well as Foundry and that you will be assessed on how well you integrate those capabilities into your demo. They also want to understand why you chose the problem, how you approached it with their products, and the solution you built.Source 1Deployment Strategist, Build to Apply - US Government (Washington, D.C.) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 2Deployment Strategist, Build to Apply - US GovernmentPublisherPalantir Technologies (Lever)Source typecompany job posting

How long should the Build to Apply video be?

Palantir’s Build to Apply postings ask for a demo video under 5 minutes, submitted as an unlisted YouTube link, within 7 to 10 days of starting to build.Source 2Deployment Strategist, Build to Apply - US GovernmentPublisherPalantir Technologies (Lever)Source typecompany job posting

Can I use ChatGPT or Claude for Palantir Build to Apply?

Yes. The postings say LLMs, AI and other technical tools are permitted and encouraged while you build the project.Source 1Deployment Strategist, Build to Apply - US Government (Washington, D.C.) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 2Deployment Strategist, Build to Apply - US GovernmentPublisherPalantir Technologies (Lever)Source typecompany job posting

Is Build to Apply the same as the Palantir FDSE application?

No. The postings are titled Deployment Strategist, Build to Apply, US Government. On Palantir’s job board the Deployment Strategist role sits under a different team from the Forward Deployed Software Engineer role.Source 1Deployment Strategist, Build to Apply - US Government (Washington, D.C.) - Palantir on LeverPublisherPalantir Technologies (Lever)Source typecompany job postingSource 7Forward Deployed Software EngineerPublisherPalantir Technologies (Lever)Source typecompany job postingSource 8Deployment StrategistPublisherPalantir Technologies (Lever)Source typecompany job postingSource 9Software Engineer - Apollo PlatformPublisherPalantir Technologies (Lever)Source typecompany job posting

What problem should I build for Palantir Build to Apply?

Pick one named user who makes one recurring decision, using public or notional data, and build the path from raw data to that decision with an AIP step inside it. A narrow workflow that runs end to end shows more than a broad dashboard. This is our method, not a rule Palantir publishes.

Keep reading

Guides

Company guides

Lessons

Questions

More in Company loops