How it works
A second perspective on statistical programming.
We’re developing Akratos to help teams understand where results differ and what needs attention.
This is the intended workflow for a product in development.
Start with the study
Use the study’s analysis plan, specifications and data as the starting point. Missing or conflicting requirements need to be resolved before the affected work proceeds.
Check independently
Develop a separate QC implementation for comparison with the study outputs. The intended separation keeps production code out of the generation inputs.
Review the findings
Bring discrepancies back to the team for investigation and approval under its own procedures.
Agreement alone does not establish correctness. The study’s requirements remain the reference for investigating a difference and deciding whether the specification, source data or either implementation needs attention.