Format bridges
CSV, Parquet, fixed-width and the occasional decades-old export format. Converters that keep types and nulls intact instead of rounding them into something friendlier.
Batch data tooling, built small.
We write narrow command-line tools for moving, validating and reshaping tabular data — the unglamorous middle of a data pipeline. No platform, no dashboard, no seat licences.
What we build
Validation passes
Schema and range checks that run before a load, not after. A failing row is reported with its line number and the rule it broke, so the fix takes minutes rather than an afternoon.
Reconciliation
Comparing two extracts that should agree and explaining precisely where they do not — by key, by column, by magnitude of difference.
Approach
- Small pieces
- One tool does one transformation and writes to standard output. Pipes are the integration layer; there is nothing else to learn.
- Deterministic runs
- The same input and the same version produce byte-identical output. That makes a diff a meaningful test, and a test a meaningful gate.
- Data stays where it is
- Everything runs on the machine holding the data. Nothing is uploaded anywhere for processing, because nothing needs to be.
Contact
Work enquiries: hello@kaskodata.net. We answer on weekdays, usually within a day.