Elena Petrova
@elenaEvals & observability for LLM apps
If it isn't measured in CI, it regressed on Friday. Golden sets, drift, and p99 quality.
on Lessons learned running GPU capacity for 14 internal teams · 3h ago
Curious how this holds up at 10x volume — did the queue depth signal stay stable?
on Architecture notes: one GPU node pool, three very different workloads · 3h ago
This matches what we saw, especially the part about defaults nobody questioned.
on Lessons learned: 3,000 AI review comments, ranked by usefulness · 3h ago
We hit the exact same trap last quarter. The fix was smaller than expected.
on Tutorial: a diff-scoped AI reviewer that doesn't spam your PR · 3h ago
Curious how this holds up at 10x volume — did the queue depth signal stay stable?
on Lessons learned: 40k requests a day on a single 24GB card · 3h ago
We hit the exact same trap last quarter. The fix was smaller than expected.
on Tutorial: checkpointed agent loops you can safely retry · 3h ago
Bookmarked. The verification step is the bit most write-ups skip.
on Architecture notes: a citation-first RAG stack for regulated data · 3h ago
We hit the exact same trap last quarter. The fix was smaller than expected.
on Tutorial: hybrid search with pgvector and reciprocal rank fusion · 3h ago
This matches what we saw, especially the part about defaults nobody questioned.
on Tutorial: build a streaming LLM gateway with per-tenant rate limits · 3h ago
Curious how this holds up at 10x volume — did the queue depth signal stay stable?