Dataloop is a data-centric AI platform built around four pillars: Data (unstructured data management, automated preprocessing, embeddings, versioning), Models (deploy, fine-tune, version, and experiment with LLMs and foundation models), Pipelines (visual and code-based orchestration of data, models, apps, and human feedback), and Applications (function-as-a-service for custom AI logic). The platform exposes a Python SDK and REST API, integrates with NVIDIA NIM for accelerated GenAI adoption, and supports multimodal workloads including LiDAR, audio, NLP, and computer vision. It is compliant with SOC 2 Type II, ISO 27001, ISO 27701, and GDPR, with RBAC, SSO, 2FA, and AES-256 encryption.