Ontology Learning Round
The brief
You've been handed a small slice of a company's Ontology: Employee and Team objects with properties, connected by typed links for who reports to whom and who is on which team. The catch is you have never seen this Ontology API before, and the only documentation is the docstrings in one read-only file. A stakeholder wants three org-chart answers out of it, and the API does not work the way you might expect.
Across five milestones you'll learn the unfamiliar Ontology API cold, decompose three questions into the primitives it actually gives you, design the one query that needs a real traversal, implement all three in the live codebase until the tests pass, then reflect on how you learned a system you did not build.
Why this company
Palantir's Learning round hands you an unfamiliar system, API, or language feature with minimal docs and watches how fast and accurately you build a working mental model, then apply it to problems of increasing difficulty. It maps directly to the forward-deployed job: an FDSE lands at a new client and has to learn their Foundry Ontology and data model cold, then be useful fast. This lab mirrors that round using Palantir's own core abstraction, the Ontology of objects and typed links. It is the gentler, intern-friendly end of the loop: the API is tiny and the org is small, but the signal is whether you read before you guess. It is one round of a longer loop, so treat it as onsite prep, not the whole bar.
What you'll do
- 1ClarifyLearning roundLearn what the unfamiliar API actually does before you write a query.
- 2DecomposeLearning roundMap the three questions onto the primitives the API really gives you.
- 3DesignLearning roundCommit to a correct, terminating traversal for the hardest query.
- 4BuildLearning roundTurn the plan into working code against the API you just learned.
- 5ReviewSelf-review + mission reflectionReflect on how you learned a system you did not build.
What this prepares — and what to prep elsewhere
This lab rehearses
- Learning / Applied Learning round (model an unfamiliar API cold, then solve problems of increasing difficulty)
- Reading minimal docs and building a mental model before writing code