In autonomous trading, the kill-switch is the architecture.
The Trading Terminal runs unattended. That is only possible because the risk caps, position sizing, and shut-off logic were designed before the alpha logic — not after.
Most trading systems begin with the strategy: a signal, an indicator, an alpha. The risk layer arrives later, bolted on, and is the first thing to fail under stress.
The Enso Trading Terminal was built backwards on purpose. The kill-switch came first. Risk caps were schemas, not config. Position sizing was a typed function with property tests, not a magic number in a YAML file.
The result is a system that can run unattended without anyone losing sleep. Strategy code can be wrong. Risk code cannot.
For client engagements building autonomous trading, the rule is the same: the risk surface is the architecture. Strategy is a plug-in.
Want to scope an engagement around this?