Articles
Notes on building signal systems.
Practical essays on data engineering, evaluation, and the decisions that make analytical systems useful in production.
-
Start with the decision
A short decision brief makes it possible to tell whether a new signal is worth building.
-
What a research result needs before production
Give the next operator enough information to reproduce, observe, and reverse the change.
-
Two clocks and a correction
How to reconstruct what happened and what the system knew when it acted.
-
The average can hide the failure
Break an evaluation into slices that match the consequences of getting a decision wrong.
-
Design the handoff for the reviewer
A useful alert carries enough evidence for the next person to act and explain why.
-
A threshold is an operational setting
Treat a threshold change as a change to queue volume, reviewer workload, and error costs.
Articles by Marina Volkova, Ilya Sokolov, Anna Lebedeva, and Daniel Mercer use fictional editorial personas.