skills.rest
LigphiDonk
LigphiDonkCOMMUNITY
Explore all skills by this author →
GitHub

inno-experiment-dev

Orchestrate ML experiment lifecycles from planning to submission with judge feedback.

708|51|Updated Mar 10, 2026
Search Tags:#planning#checkpoints#code-generation#experiment#ml#submission#judge-feedback
One-click install
npx skills add https://github.com/LigphiDonk/Oh-my--paper --skill inno-experiment-dev-ligphidonk
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: inno-experiment-dev
Source: https://github.com/LigphiDonk/Oh-my--paper/tree/main/src-tauri/resources/skills/inno-experiment-dev
Command: npx skills add https://github.com/LigphiDonk/Oh-my--paper --skill inno-experiment-dev-ligphidonk

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill automates the end-to-end lifecycle of ML experiments, from upfront planning through code implementation with a judge feedback loop to final submission, reducing manual coordination and errors.

Core Features & Use Cases

  • Automated planning and orchestration: Generates a detailed implementation plan mapping datasets, models, training, and evaluation, bridging ideas to executable code.
  • Code writing with feedback loop: Implements the project following the plan while iteratively incorporating Judge agent feedback to improve quality and correctness.
  • Experiment submission and artifacts: Runs the experiment (with GPU support), logs results, saves final checkpoints, and submits the final run for analysis.

Quick Start

Run the end-to-end ML experiment workflow: generate an implementation plan, implement the project with judge feedback, and submit the final experiment run.

Frequently Asked Questions about inno-experiment-dev

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate the end-to-end ML experiment lifecycle from planning to submission?▼

Automating the end-to-end ML experiment lifecycle requires orchestrating planning, code implementation, and evaluation through a structured workflow. This process coordinates agents to generate implementation plans, write code with iterative judge feedback, and submit final experiment runs.

What is the best way to incorporate judge feedback into machine learning code generation?▼

Incorporating judge feedback into ML code generation involves an iterative loop where a Judge Agent evaluates the implementation. The system refines project code based on this feedback to improve quality and correctness before final submission.

How do I manage ML experiment checkpoints and ensure GPU-backed training?▼

Managing ML experiment checkpoints requires a workflow that enforces GPU-backed training and saves model states to a designated path like `model_final.pth`. The system handles artifact management by logging results thoroughly during the experiment run.

Do I need to review a plan before orchestrating ML code generation and evaluation?▼

Reviewing a plan before orchestrating ML code generation is required. The workflow is designed for use after code-survey and plan review to properly orchestrate the Coding Plan Agent, ML Agent, and Judge Agent through the experiment lifecycle.

Related Skills

Explore More Software Engineering Skills→
AngelP17AngelP17
1

ml-pipeline-workflow

Automate end-to-end ML pipelines from data ingestion to deployment.

Community
Advanced
bglick13bglick13
1

experiment-analysis

Diagnose GRPO training runs using WandB metrics and Axiom logs.

Community
Advanced
databricks-solutionsdatabricks-solutions
1.8k

mlflow-evaluation

Automate GenAI evaluation of ML agents with MLflow workflows.

Official
Advanced
enigmaicon-engenigmaicon-eng
1

ml-pipeline

Generate production ML training pipelines with MLflow logging and Kubeflow or Airflow orchestration.

Community
Advanced