Term tag: AI Architecture
AI Architecture
How a production AI system is assembled: retrieval, context assembly, the model call, tool execution, validation, logging and evaluation.
Mixture-of-Experts (MoE)
The architecture behind "671B total, 37B active": hundreds of expert sub-networks, a router that picks a few per token, and a memory bill that never shrinks.
Model Routing
Model routing sends each request to the cheapest capable model — a small model handles the easy majority and hard queries escalate to a frontier model.
Residual Connections (Skip Connections)
Residual connections add a block's input back to its output — F(x) + x — the trick that lets gradients flow through deep nets and made transformers trainable.