Field Guide
Forward-Deployed AI Engineering for AEC.
Definitions, architecture and working patterns for applying AI inside real AEC and engineering workflows — each one tied back to something actually built, not offered on its own.
Read in this order, or jump to what you need
Definitions, architecture, patterns, implementation.
What forward-deployed engineering actually means, why it matters more for AI than for typical software, and the Stack framework that formalizes it.
Read →Why AEC is a different problem for AI than most software domains, and the investigation/determination boundary that makes a result trustworthy.
Read →Physical problem → formal model → intelligent system → verified result → production product — the one shape every project on this site follows, plus the interactive architecture canvas that puts it into practice.
Read →What it takes to ship a Revit add-in as a production system: host lifecycle, multi-version compatibility, deterministic QA and release validation.
Read →Framework
The AEC Forward-Deployed Stack
Seven layers, domain to outcomes, with the field engineer's work cutting across all of them. The first canonical diagram in this field guide — more will follow as the practice gets written down.
Proof, not just argument
Every idea in this field guide traces back to a system that actually shipped.