Skip to content

Cloud & AI security · Zürich area

Cloud systems that hold up under real decisions.

Artem Kozlenkov advises and builds across Azure, AWS and GCP — modernization, platform engineering, governance, and agentic AI with security boundaries that actually hold.

  • Zürich area
  • Azure · AWS · GCP
  • 7+ engineers coordinated
  • Applied AI security
Three stacked architecture planes for identity, network and workload, with one clay path marking an enforced boundary.
Fig. 01 — Reference architecture: identity, network, workload.

Capabilities

Four ways I get teams unstuck.

All services

How the work runs

Small steps. Visible trade-offs. No theatre.

Architecture decisions are only useful when the teams who inherit them can see why they were made — and change them safely.

A

Start from the constraint that hurts

Migration, audit, delivery speed or AI risk — we name the real constraint before proposing a platform.

B

Make the boundary explicit

Identity, network and workload boundaries are documented as decisions, not left to tribal knowledge.

C

Leave the team able to operate it

Runbooks, guardrails and a paved path that survives after the engagement ends.

Start a conversation

Bring a hard platform problem.

Share where you are, what is in the way, and what a useful next step looks like. You do not need a finished brief.

Reply time
Within two business days.

Your details are used only to reply. Privacy.