Series

Five threads run through the site. Each page lists its pieces in the order they were written.

• Small data 0

  Machine learning on tens of runs: validation, what not to predict, and instrumentation decided before the campaign.

• Private by default 0

  Local models and retrieval for work that cannot leave the building.

• Tested on real work 6

  An assistant given a real engineering task, and where it failed.

• From zero 11

  A PhD engineer learning Python and machine learning: what to learn first and what to skip.

• What a pilot is for 0

  Scale-up method and judgment, as general lessons.