A
89.0Overall score
A data and ML setup that explores a dataset, engineers features, and tracks every run with seeds and metrics so results stay trustworthy. For practitioners who want an agent that guards against leakage instead of overstating accuracy.
89.0Score
5Components
Install this build
terminal
pip install jupyterlab mlflow && jupyter labComponents
Model
- Claude Opus 4.8
MCP servers
- jupyter
- postgres
- filesystem
- github
Subagents
- eda-explorer
- feature-engineer
- eval-reviewer
Hooks
- PostToolUse: run cell and capture metrics
- Stop: log experiment to tracker
Rules
- Seed everything for reproducibility
- Never leak the test set into features
Summary
A data and ML setup that explores a dataset, engineers features, and tracks every run with seeds and metrics so results stay trustworthy. For practitioners who want an agent that guards against leakage instead of overstating accuracy.
89.0 score
0 Reviews
Your rating
Sign in to post
Loading discussion...