legit-embedding
operator feed · project room
legit-embedding feed
Persisted project activity, filtered so routine daemon check-ins stay out of the way.
6 hidden
working set
Active refs
6
lane
Needs you
0
Nothing needs you
Open questions, MIA daemons, pending feedback, and compaction requests will land here.
stream
Project feed
19 shown
- Brief #206: Now depends on #205 (Build single-process multi-model async worker + worker-side reply_to (Track B)). Activity brief_events · dependency_added · conductor-claude · 1d ago
- Brief #205: Now depends on #204 (Spike: choose the single-process multi-model serving substrate (Track B)). Activity brief_events · dependency_added · conductor-claude · 1d ago
- Brief #207: Grouped under epic #201. Activity brief_events · parent_set · conductor-claude · 1d ago
- Brief #206: Grouped under epic #201. Activity brief_events · parent_set · conductor-claude · 1d ago
- Brief #205: Grouped under epic #201. Activity brief_events · parent_set · conductor-claude · 1d ago
- Brief #204: Grouped under epic #201. Activity brief_events · parent_set · conductor-claude · 1d ago
- Brief #202: Grouped under epic #201. Activity brief_events · parent_set · conductor-claude · 1d ago
- Brief #207: # Text-embedding path: first live validation Repos: **lounge** (turn on) + **conductor-client** (consumer) + **legit-embedding** (worker). Was Solo todo 664. **Independent** — runn... Activity brief_events · plan_proposed · conductor-claude · 1d ago
- Brief #206: # Worker-type VRAM cost table + greedy packer (Track B) Repo: **legit-embedding**, on the #205 single-process runtime. **Depends on #205.** This is the operator's "dynamic memory /... Activity brief_events · plan_proposed · conductor-claude · 1d ago
- Brief #205: # Build single-process multi-model async worker + worker-side reply_to (Track B) Repo: **legit-embedding**, built in an **ISOLATED worktree** (worktree-manager) + its own **Solo pr... Activity brief_events · plan_proposed · conductor-claude · 1d ago
- Brief #204: # Spike: choose the single-process multi-model serving substrate (Track B) Repo: **legit-embedding** (isolated worktree). A decision gate that feeds #205. Latency-insensitive workl... Activity brief_events · plan_proposed · conductor-claude · 1d ago
- Brief #202: # Image batch VRAM auto-sizing on the current worker (interim throughput win) Repo: **legit-embedding** (branch `runpod-container-runtime`). Conductor companion config already done... Activity brief_events · plan_proposed · conductor-claude · 1d ago
- Brief #201: # Epic: Conductor multi-model + multi-app embedding (the vodmanager foundation) **Objective:** scale conductor's GPU embedding pipeline for **throughput-per-dollar** on large defer... Activity brief_events · plan_proposed · conductor-claude · 1d ago
- Brief #207: The text wire-format + pipelining shipped but text has never run live (was Solo todo 664). Cheap single-app end-to-end validation on the current worker; also the natural first seco... Activity brief_events · note_added · conductor-claude · 1d ago
- Brief #206: Per-worker-type resource table (base_mb, per_item_mb, batch bounds, in_flight, target_depth) + a GREEDY VRAM packer that fits a heterogeneous model mix weighted by queue depth, com... Activity brief_events · note_added · conductor-claude · 1d ago
- Brief #205: Rewrite the worker to ONE process that loads each model once and hides the slow XADD behind async concurrency instead of duplicate model processes. Includes the worker-side reply_t... Activity brief_events · note_added · conductor-claude · 1d ago
- Brief #204: Decide the substrate for the single-process multi-model async worker: Ray Serve vs custom asyncio (Triton likely too rigid). Prototype must prove load-once + async xadd/compute ove... Activity brief_events · note_added · conductor-claude · 1d ago
- Brief #202: Auto-size IMAGE_BATCH_SIZE to the pod's FREE VRAM (nvidia-smi, resolved pre-Popen in start_workers.py) so big cards run ~80+ instead of pinned 32 — ~2.5x on the live image backfill... Activity brief_events · note_added · conductor-claude · 1d ago
- Brief #201: Umbrella epic grouping the conductor GPU-embedding throughput + multi-app work: two tracks (Laravel-side result-routing contract now; isolated single-process multi-model worker rew... Activity brief_events · note_added · conductor-claude · 1d ago