Run your data warehouse end-to-end.
Best-practice guides organized by role. Each one teaches the practice first, then shows how Kirimana operationalizes it. Pick your role to see the track built for you.
The catalogue is open; reading a guide requires a (free) account. Sign in or request access .
12 guides in this track · updated with every release
Security & AI policy
- Put a gate on every AI call to your data 🔒
AI access is an attack surface. Route every model call through one audited gateway that classification-gates it and refuses by default, because you cannot secure what bypasses the gate.
- Classify data so protection is automatic 🔒
Make classification the control plane. When every column carries a sensitivity label, masking, AI refusal, and access follow automatically, and personal data can never silently become public.
- Control access with single sign-on and capability-based roles 🔒
Pin identity to your identity provider, scope capabilities per environment, and enforce them at pull-request time so a risky change is stopped before it merges, not after.
- Keep secrets out of code and configuration, provably 🔒
No plaintext secret in code or YAML, ever. Reference secrets by vault token, resolve them at runtime, and fail the build on any plaintext that slips through.
- Gate natural-language questions before they run 🔒
A question asked in plain language over your data is a new exfiltration path. Resolve each table to its contract and each column to its classification, and decide allowed, gated, or forbidden before the query runs.
Compliance & audit
- Meet your GDPR obligations in a contract-first warehouse 🔒
Turn data minimization, lawful basis, subject rights, and retention limits into contract metadata a validator enforces, instead of a policy binder nobody reads.
- Redact for the right to be forgotten without destroying the audit trail 🔒
Resolve the apparent conflict between erasure and auditability with redaction-as-event, a two-person countersignature, and a clock that starts when you acknowledge the request.
- Prepare for the EU AI Act as a data platform 🔒
Know where AI touches your data, under what policy, with what audit trail, and produce structured evidence a reviewer can attest to.
- Build a continuous evidence base for audits 🔒
Generate compliance evidence from live metadata and the audit log over a defined control set, store it where your GRC tooling can import it, and keep a history instead of scrambling before each audit.
- Meet operational-resilience expectations across regulations 🔒
Treat detection, incident handling, and evidence as one resilience practice that satisfies DORA in finance and equivalent operational-risk expectations everywhere else, alongside EU AI Act and GDPR duties.
DataOps
- Run a data incident like a security incident 🔒
Classify severity, communicate, contain, and run a blameless postmortem with lineage as evidence, because a data-quality breach and a data exposure share one playbook.
- Route detection events to your service desk with regulator-grade trace links 🔒
Detection is worthless if it dies in a log. Route every event to the system your responders already use, carrying the trace links a regulator will ask for, so every alert is actionable.