scikit-learn

For numeric targets I usually start simple and make regularization earn its keep. Ridge is stable, Lasso helps with sparsity, and ElasticNet is a practical compromise when correlated features exist. The main goal is not just minimizing RMSE but unders

Categoricals are where good intentions become leakage. I use one-hot encoding for low-cardinality stable fields, ordinal encoders only when order is real, and frequency or target encoders with strict cross-validation boundaries. The encoder strategy s

Hyperparameter search should be targeted, not theatrical. I usually combine a strong baseline, a compact search space, and a metric aligned with business cost. GridSearchCV is good for interpretable sweeps; randomized search is better when the space g

Deployment should not rewrite the feature logic from scratch. I expose trained pipelines behind FastAPI so the exact preprocessing and estimator objects travel together. Strong request schemas and explicit model versioning keep this boring in the righ

Before I fine-tune transformers, I almost always try a TF-IDF baseline. It is fast, interpretable, and often surprisingly competitive for moderate text classification tasks. If a linear model over sparse features is already good enough, that is usuall

I push nearly all preprocessing into a Pipeline so training and inference paths share exactly the same logic. ColumnTransformer is the workhorse here because real-world tables mix numeric, categorical, boolean, and text fields. It gives you reproducib

Evaluation goes wrong when data splitting is treated like boilerplate. I stratify imbalanced targets, guard time order when necessary, and make sure preprocessing lives inside cross-validation. This is the difference between a model that looks good in

I like setting a few strong baselines before chasing complexity. A regularized logistic regression, a random forest, and a gradient boosting model usually tell me whether the problem is linearly separable, non-linear, or data-limited. Good baseline di