Datadog and Baseten fine-tune Qwen 3.5 9B for a 50x faster LLM summarizer
baseten · x · 2026-11-19
Baseten partnered with Datadog to make the Patterns LLM topic-clustering summarizer 50x faster by fine-tuning Qwen 3.5 9B on production traces: latency dropped from 44.7s to 0.89s and judge pass rate rose from 45% to 93%.
They open-sourced pup-finetune (Apache-2.0), which curates datasets from Datadog Agent Observability trajectories (messages, tool calls, results), fine-tunes an open-weight model, and evaluates it against the base model — with a datadog-finetune skill so your coding agent can run the whole workflow.
More from coding & agent
- Developer adds a real-time talking-head avatar to his local voice agent using open-weight video model — Scobleizer · 2026-11-19
- Personal agent auto-clips a livestream into highlight videos with just two prompts — Scobleizer · 2026-11-19
- Beyond a single bot: building a small agent fleet with OpenClaw Specialist Teams — heyneighbor · 2026-11-19
- Cognition's Walden Yan gives away 10,000 Devin Bots with up to $10K credits via Text Poke — marvinvonhagen · 2026-11-19
- mitsuhiko's agent verifies AirPlay hack at 3 AM by blasting sound through the living room mic — altryne · 2026-11-19
- mitsuhiko's overnight agent woke him at 3AM blasting audio to verify AirPlay reverse engineering — mitsuhiko · 2026-11-19