agentic-kaggle-skill

Agentic Kaggle Skill

Operating Loop

Treat every competition as a validation problem first and a modeling problem second. The default target platform is Kaggle, so prefer Kaggle-native notebooks/scripts, datasets, model artifacts, competition submissions, and score receipts. For code competitions, assume the final notebook/kernel will be rerun by Kaggle against hidden data unless the competition docs prove otherwise.

1. Read the competition page, classify the submission mode as classic file submission or code/notebook scoring, then inspect the rules, data-use terms, sharing policy, data dictionary, metric, submission format, train/test construction hints, and leakage warnings.
2. If live competition intelligence tools are available, inspect top public open notebook solutions and relevant discussion activity before major architecture choices; treat them as clues, not authority.
3. Identify the task type: binary, multiclass, multilabel, regression, ranking, image, segmentation, text, time series, grouped entities, or a hybrid.
4. Design folds before feature engineering or modeling. Prefer a fold column saved into the training data so every experiment uses the same comparison surface.
5. Build the simplest metric-correct baseline and produce out-of-fold (OOF) predictions plus a valid submission.
6. Proactively plan a stronger architecture once the baseline is trustworthy: diverse model families, feature/embedding producers, augmentation/pseudo-label/distillation stages, calibration/postprocessing, and an ensemble or stacker.
7. Iterate with validation gates: test one meaningful change at a time when possible, but launch several independent producer notebooks in parallel when they create diverse artifacts that can be compared by OOF score or ensemble diversity.
8. Offload heavy training, inference, embedding generation, image/text experiments, or memory-risky jobs to Kaggle notebooks/scripts when local compute may OOM or take too long.
9. For sophisticated architectures, split work into a small Kaggle pipeline: run several independent producer notebooks/scripts first, save each useful producer output as a private Kaggle dataset, then run one consumer notebook/script that attaches those datasets and creates the final OOF/test/submission outputs. When a producer trains a model, its checkpoint, tokenizer/config, fold metadata, OOF/test predictions, and manifest should be exported as a Kaggle dataset; downstream notebooks load the model from /kaggle/input/....
10. Submit the final submission-producing artifact to Kaggle for scoring and retrieve the resulting submission status/score.
11. If Kaggle returns a code-competition error or vague scoring failure, enter the debugging loop: retrieve available logs, classify likely failure mode, patch defensively, rerun the final kernel, resubmit, and repeat until scored or concretely blocked.
12. Track local CV, remote Kaggle run status, Kaggle submission score, public LB, private-risk notes, seed, code version, data version, and artifact paths for every run.
13. Ensemble only with OOF predictions generated without in-fold leakage.