The Palantir Ontology session was hands-down the clearest walkthrough of Object Types and Actions I've seen. It saved me a month of trial and error at work.
ai solutions
Intelligence, hand-built.
From GPU-native inference to agentic systems — the essence of what we compose, painted the way we build: with intent, texture, and craft.
GPU inference · NVIDIA · AMD
GPU-Native Inference
vLLM, PagedAttention, and quantization — right-sized serving tuned for NVIDIA and AMD accelerators at real cost.
data platform
Palantir Foundry
Ontology, medallion pipelines, Workshop apps, AIP agents, and OSDK — practitioner-grade delivery and enablement.
knowledge
RAG Systems
Retrieval-augmented pipelines that ground models in your own trusted knowledge — accurate, cited, and current.
models
LLMs & SLMs
Fine-tuned language models — from frontier LLMs to compact, cost-efficient SLMs serving in under 100ms.
autonomy
Agentic Systems
Orchestrated AI agents that plan, call tools, and complete real work — with guardrails and human oversight.
bespoke
Custom AI Products
End-to-end AI products — like our GMAT AI tutor — designed, built, and operated in-house.