Document AI Inference architecture and freeze current baseline
This commit is contained in:
@@ -2,6 +2,13 @@
|
||||
|
||||
Date: 2026-09-03. Owner-approved target; the first local modular vertical slice is installed.
|
||||
|
||||
Current implementation snapshot (2026-09-05): [AI Inference architecture](../audits/2026-09-05-ai-inference-architecture.md).
|
||||
The owner has temporarily frozen this stage; user acceptance remains incomplete.
|
||||
This ADR describes the target boundaries. The snapshot explicitly distinguishes
|
||||
the implemented setup mapping and source/result reuse from the not-yet-wired
|
||||
general DAG executor and per-node cache. The product name is now AI Inference;
|
||||
technical Observatory identifiers remain compatible.
|
||||
|
||||
## Decision and supersession
|
||||
|
||||
A reusable module is an independently packaged Docker image for a model or a
|
||||
|
||||
Reference in New Issue
Block a user