Orchestrate polars-native ML lifecycles with ONNX export and registry governance.
npx skills add https://github.com/FFOO6866/lead2cash --skill kailash-ml
Please help me install this Agent Skill. Skill: kailash-ml Source: https://github.com/FFOO6866/lead2cash/tree/main/.claude/skills/34-kailash-ml Command: npx skills add https://github.com/FFOO6866/lead2cash --skill kailash-ml
Kailash ML solves the complexity of building reliable end-to-end machine learning lifecycles—from feature ingestion and training to registration, ONNX serving, and drift monitoring—while enforcing safety and consistency constraints that prevent common production failures.
Train and register a production-ready model by running an async pipeline that tracks a run, trains from a polars DataFrame, then registers and serves the resulting artifact using the engine-first km.* verbs.
High-intent search queries and answers about installing and using this skill.
You can build a polars-native ML pipeline by running an async workflow that tracks a run, trains directly from a polars DataFrame, and registers the resulting artifact, converting data only at framework boundaries to reduce mismatch risks.
The best way to manage a model registry lifecycle is to enforce controlled stage transitions—staging to shadow to production to archived—while applying SHA256 artifact verification and optional agent guardrails to maintain strict governance.
Export sklearn, lightGBM, or PyTorch models to ONNX by running the export pipeline that verifies outputs before deployment, enabling cross-language compatibility and serving via an ONNX-first runtime through Nexus.
Yes, asynchronous pipelines support AutoML agents by providing optional agent guardrails with human approval, enforcing safety and integrity constraints during the automated training and model registration lifecycle.
Drift monitoring triggers retraining hooks by watching for data drift within the lifecycle, automatically initiating the asynchronous resuming and training stages to update and redeploy the model artifact.
SHA256 artifact verification is required for controlled stage transitions to enforce production-grade governance, ensuring integrity checks pass before a model moves between staging, shadow, and production environments.
1thirteeng3Guide RL-based post-training of GLM models using slime with Megatron-LM integration.
nirukk52Monitor Encore.ts and SvelteKit logs to diagnose drift and verify run events.
doanchienthangdevDeploy machine learning models with TensorFlow Serving, TorchServe, or Triton Inference Server.
terrene-foundationOrchestrate ML training, registry, monitoring, and ONNX serving via km.* APIs.
