A method worth keeping close.

Andrej KarpathyAnalytical perspectiveAK-P02

PRODUCT ADVISORY COUNCIL

Understand data before complexity

Data and preprocessing errors may matter more than model structure.

Andrej KarpathyAnalytical perspectiveAK-P02

The method

Understand data before complexity

Data and preprocessing errors may matter more than model structure.

When to use it

• Unclear results in a data-driven product or model.

Ask yourself

• Poor results: should we build something more complex?

Put it into action

• Inspect real and edge-case samples, labels, duplicates, damage, and bias. Confirm what actually enters the system before adding complexity.

Keep these limits in mind

• This adapts a training practice to product work; more data is not automatically better.

An analytical method adapted from the public skill package, not a reply or endorsement by the person represented.
View the public method source ↗