Live release desk / KernelMark / Sliema

No model ships without a record.

Inventory live Eval live Approvals live Cost live Scroll to enter
KernelMark private beta for EU AI product and ML platform teams Book a readiness session Inventory, eval history, approvals, and spend in one governed workspace
EU-first model governance workspace

Know every AI system before it becomes a risk.

KernelMark gives AI product, ML, and compliance teams one living workspace for model inventory, evaluation history, ownership, approval status, and cost.

Request demo Explore platform
0 AI assets tracked
0 approval coverage
0 cost anomalies found
app.kernelmark.com / inventory
Live sync
AI SYSTEM REGISTER

Production AI systems

Search models, owners, evals...
Needs owner 14 3 critical
Eval gate pass 91% +7% this month
July AI spend EUR 42.8k 6 vendor caps
SystemOwnerLast evalStatus
Credit Risk CopilotRisk Opsmx-eval-1429Approved
Claims AssistantProduct AImx-eval-1392Review
Fraud Graph AgentUnassignedMissingBlocked
Eval release gate

Regression suite attached to version v4.2 before approval.

Spend anomaly
+38%

Vendor usage exceeded finance policy threshold.

Connects to the stack AI teams already use

MLflow Databricks AWS Bedrock Azure AI Google Vertex AI Hugging Face Jira GitHub Slack Snowflake MLflow Databricks AWS Bedrock Azure AI Google Vertex AI Hugging Face Jira GitHub Slack Snowflake
0

AI systems tracked across design partners

0

average minutes saved per release review

0

EU evidence fields mapped per high-impact system

0

monthly spend anomalies surfaced before approval

The operating gap

AI shipped faster than governance workflows could keep up.

Can you list every AI system in production?
Who approved the last model version?
Which models are over budget?
Which systems need EU AI Act evidence?
One architecture, end to end

From scattered models to governed AI systems.

KernelMark connects source systems, evaluation evidence, ownership, approval workflows, and cost telemetry so every AI system has a clear operating record.

Inputs
Model sources
Product workflows
Spend systems
Governance layer
KernelMark control plane Inventory, Eval, Cost, approvals, and evidence
Outputs
Approval status
Audit evidence
Cost control
EU readiness
Why KernelMark

Govern every model from intake to approval.

01

Inventory that stays current

Auto-sync AI systems from ML platforms, cloud providers, vendor tools, and internal registries.

02

Eval evidence attached to releases

Store test suites, regression runs, safety checks, and reviewer decisions beside every model version.

03

Approval workflows for AI systems

Route high-impact changes through owners, compliance, security, and product leaders with a clean audit trail.

04

Cost visibility by model and vendor

Track usage, budget caps, vendor drift, and chargeback-ready reporting for AI spend.

05

EU-ready reporting

Export inventory, ownership, evaluation, DPIA, and approval evidence for internal reviews and regulators.

Core products

Inventory, Eval, and Cost are designed to work together.

KernelMark Inventory

Register AI systems, owners, data domains, model versions, vendors, risk tier, and deployment status.

mx-credit-risk-v4Approved
claims-assist-prodNeeds review
fraud-agent-betaOwner missing

KernelMark Eval

Keep release gates, evaluation history, regression results, prompt tests, and reviewer notes in one evidence trail.

Bias suite passed

Latency regression flagged

Risk owner requested review

KernelMark Cost

Monitor AI spend by vendor, model, business owner, product, and policy threshold before overruns spread.

Live control room

Governance activity your teams can actually act on.

The workspace turns model events into readable ownership, evaluation, cost, and approval signals.

ASK KERNELMARK
Model approvedCredit Risk v4 passed Eval Gate
Cost anomalyGPT vendor spend up 38% this week
Owner assignedFraud Assist now owned by Risk Ops
Evidence exportedEU AI Act inventory report generated
Policy matchedISO 42001 control linked to eval evidence
Built for the teams who govern AI

One workspace, different operating views.

AI product teams

Ship model updates with clear owner, release, eval, and approval status.

ML platform teams

Connect registry events, deployment metadata, and eval runs into a shared system record.

Risk and compliance

Prove which models exist, who approved them, and what evidence supports the decision.

Regulated enterprises

Run consistent controls across departments, vendors, high-risk use cases, and regions.

Trust and compliance

Evidence for the frameworks AI teams are already asked about.

EU

EU AI Act readiness

Inventory, intended purpose, human oversight, risk tier, and evidence exports.

ISO

ISO/IEC 42001

AI management system controls mapped to ownership and review workflows.

NIST

NIST AI RMF

Govern, map, measure, and manage signals connected to model records.

GDPR

GDPR and DPIA support

Data domain, processor, purpose, and risk review context for AI systems.

Pricing

Start with inventory. Expand into eval and cost control.

Starter

EUR 790/month

For AI product teams formalizing their first model inventory.

  • Up to 50 AI systems
  • Ownership and approval workflow
  • Basic eval evidence history
  • Email support
Talk to sales
Most popular Growth

EUR 2,400/month

For ML and governance teams managing multiple products and vendors.

  • Up to 300 AI systems
  • Inventory, Eval, and Cost modules
  • SSO, audit exports, and Slack/Jira workflows
  • EU readiness reporting
Request demo
Enterprise

Custom

For regulated companies with complex AI governance and residency needs.

  • Unlimited AI systems
  • Dedicated environment
  • Custom controls and policies
  • Implementation partner support
Talk to sales
Design partner signals

Built with the questions real AI teams ask during release review.

"

We stopped treating AI governance like a spreadsheet project. KernelMark gave product, platform, and risk one record for every model decision.

Anna Kowalska Head of AI Governance, Sliema design partner
"

Eval evidence is no longer buried in tickets and notebooks.

Marek Nowak ML Platform Lead
"

Cost review finally has model owners, usage context, and approval history in the same view.

Eliza Zielinska AI Product Operations
Funding announcement

KernelMark secures $650K in funding from Gama VC.

KernelMark is part of Gama VC's portfolio of companies building AI model operations for complex operating environments.

View Gama VC Read announcement
Resources

Guides for teams moving from AI experiments to AI operations.

Search readiness guides, policies, and checklists
Guide

EU AI Act inventory checklist

What every high-impact AI system record should include before review.

Template

Evaluation release checklist

A practical gate for model updates, prompt changes, and vendor AI rollouts.

Playbook

AI spend governance guide

How finance and AI teams set vendor caps, chargeback, and anomaly workflows.

Company

KernelMark is based in Malta for EU teams deploying AI responsibly.

We are working with AI product, ML platform, and risk teams that need a calmer operating layer for model governance.

SliemaEU operating base
3 layersInventory, Eval, Cost
24hDesign-partner response window
Trust Center KernelMark sp. z o.o.
EU data residency roadmapReady
AI evidence exportsReady
Approval audit trailLive
ISO 42001 mappingMapped
Evidence freshness4m ago
Reports exported24
Owners assigned96%