Per-user freshness SLAs for features #61

New Issue

alvis · 2026-04-13T14:35:37Z

alvis commented

2026-04-13 14:35:37 +00:00

Document, per feature, whether it's computed just-in-time (freshness-critical: hour-of-day, calendar-next-event) or from the online feature store (batched: 7-day average completion rate). Add a feature-spec block to each ml/features/*.py that declares freshness, source, and fallback.

Document, per feature, whether it's computed just-in-time (freshness-critical: hour-of-day, calendar-next-event) or from the online feature store (batched: 7-day average completion rate). Add a feature-spec block to each `ml/features/*.py` that declares freshness, source, and fallback.

alvis added this to the M1 — Real signal milestone 2026-04-13 14:35:37 +00:00

alvis added the platform label 2026-04-13 14:35:37 +00:00

alvis modified the milestone from M1 — Real signal to M2 — AI tips + multi-source signals

2026-04-16 15:23:27 +00:00

alvis closed this issue

2026-04-25 17:03:19 +00:00

alvis referenced this issue from a commit

2026-04-26 12:08:52 +00:00

feat(features): per-feature freshness spec — JIT vs batched (#61)

alvis referenced this issue from a commit

2026-05-06 06:55:01 +00:00

docs: update CLAUDE.md with rich per-agent inference completions (#112–#116)