numerainumeraiOfficial·5 Agent Skills Included

example-scripts

AI research agent for building and deploying Numerai tournament models

Designs, runs, and reports machine learning experiments for the Numerai data science tournaments. Implements new model types, sweeps hyperparameters, and tracks BMC and correlation metrics against official benchmarks. Packages trained models into validated pickle files and deploys them to Numerai for automated weekly submissions. Eliminates manual experiment tracking, environment mismatches, and upload debugging.
npx skills add numerai/example-scripts --all -g -y
Available:

Instructs the AI agent on how to route Numerai research, model-building, and deployment requests to the correct skills, datasets, and MCP tools in this repository.

All Skills in This Repository (5)

Pure Emerald Level Indicators

Frequently Asked Questions

FAQPage Schema
How to install Numerai example-scripts?â–¼

Run `npx skills add numerai/example-scripts --all -g -y` in your terminal to install all skills in this suite globally.

How to run Numerai experiments with an AI agent?â–¼

The numerai-experiment-design skill plans configs, runs training rounds on downsampled data, and tracks BMC metrics until results plateau. Your agent then scales the winning model to full data automatically.

How to upload a model to the Numerai tournament?â–¼

The numerai-model-upload skill builds a self-contained pickle with a predict function, tests it in Numerai's Docker container, and deploys it via the Numerai MCP server.

Does this work with Codex CLI and Claude Code?â–¼

Yes. All skills follow the standard SKILL.md format and run in Codex CLI, Claude Code, and other compatible AI coding agents.

Do I need data science experience to use this?â–¼

No. You describe your model idea in plain English and the agent handles experiment design, training, reporting, and deployment for you.

Related Repositories in Data & Analytics

View All in Data & Analytics→