RedNote Machine Learning Team
Official@rednote-machine-learning
小红书机器学习引擎框架团队，RedNote Machine Learning Framework Team
Agent Skills by RedNote Machine Learning Team
Showing 1 vetted skills indexed across 1 GitHub repositories.
Frequently Asked Questions About RedNote Machine Learning Team
FAQPage SchemaWhat tasks can I perform with RedNote's inference-autopilot skill?▼
You can analyze, benchmark, diagnose, and optimize large-model inference deployments. It tunes SGLang launch parameters, runs bounded single-host GPU experiments, plans deployment topology, inspects GPU/CPU profiles, and identifies scheduler, KV-cache, communication, or kernel bottlenecks.
Who should use the inference-autopilot skill?▼
ML platform and inference engineers operating large-model serving stacks, especially those running SGLang on GPU hosts. It suits practitioners responsible for hitting latency or throughput SLOs and diagnosing performance regressions in production inference deployments.
What inputs does inference-autopilot need to optimize a deployment?▼
It works from hardware inventory, model details, workload traces, and latency or throughput SLOs. Given these inputs, it proposes operator optimizations and validates that candidate configurations improve performance without correctness or SLO regressions.
Does inference-autopilot run large-scale multi-node experiments?▼
No. Experiments are bounded to single-host GPU runs by design. It focuses on controlled benchmarking and profiling on one host, plus deployment topology planning, rather than distributed multi-node experimentation.