Teams
Data Science owns

Every model and eval harness at Veho has to survive contact with real routes, real drivers, and real deadlines. Data Science is the org that proves a change is actually better before it ever reaches production.

PythonOR-ToolsSnowflakeMLflowdbtJupyter
  • How do we build eval harnesses that catch a regression before a driver does?
  • What's the right bar for 'good enough to ship' when the baseline is already live?
This team's first post is being written.