← Return to Safe passage
Systems

Kraken Technologies / 2025–present

Safe passage.

Keeping invalid data out, and helping the right people in.

  • Django
  • Data integrity
  • OIDC

01 / Context

The problem

In electricity settlement, invalid updates and inconsistent meter data can affect downstream systems and billing. Access also needs to work across organisations with different identity providers.

02 / Responsibility

My contribution

I added validation before downstream submission, repaired migration-related data issues, and implemented and tested multi-tenant single sign-on.

03 / Approach

The decisions

Validate before submitting downstream rather than relying on later repair. For existing migration-related records, bulk repair tooling included a dry-run mode. For authentication, a custom django-allauth provider combined OIDC and Okta with automatic email-domain routing.

  1. 01

    Validate

    Block invalid updates before downstream submission.

  2. 02

    Repair

    Correct overlapping meter dates and preserve required fields; provide dry-run tooling.

  3. 03

    Verify

    Add regression tests and diagnostic logging for missing timestamps.

04 / Result

The outcome

Validation blocked approximately 483,000 invalid updates per month across two enterprise clients. Meter-date and required-field fixes prevented migration-related billing failures. A missing-timestamp fix addressed a defect responsible for 3.7 million application errors, with regression tests and diagnostic logging added. Multi-tenant SSO was implemented and tested for two organisations.

05 / Sources & context

These are separate contributions documented in the CV; the error count is the recorded impact of the defect, not a monthly saving. Internal logs, customer information and authentication configuration are not public.

Source: Harvey Bellini’s CV ↓