Industry

A map of the Generative AI industry — the stack it's built on and the roles that run it

The Generative AI stack

Five layers, from the apps people touch down to the chips everything runs on.

End users
1

Application

Most gen-media roles live here

End-user products and experiences built on top of AI models — creative tools, copilots, and generative-media apps.

Creative toolsCopilotsGen-media apps
2

Orchestration

The frameworks, agents, and tooling that coordinate models, data, and workflows — RAG, agent frameworks, and production pipelines like ComfyUI.

AgentsRAGPipelines
3

Foundation models

The large pretrained generative models themselves — diffusion, LLMs, and video/3D models — built by frontier labs and model providers.

DiffusionLLMsVideo & 3D
4

Cloud

The compute platforms and infrastructure that train and serve models at scale — hyperscalers, GPU clouds, and inference providers.

GPU cloudsInferenceTraining
5

Semiconductors

The chips — GPUs and AI accelerators — that everything else runs on. The physical foundation of the entire stack.

GPUsAccelerators
Hardware

Role categories

Most gen-media companies build at the Application layer — and hire across these six kinds of roles. Pick one to browse its tracked jobs.