Greenfield landing zones
A secure, multi-account AWS foundation built from scratch — networking, identity, guardrails, and cost controls. The boring-but-critical base your product sits on, done to industry best practice from day one.
Cloud infrastructure · DevOps · SRE
Greenfield cloud infrastructure, built right the first time.
We take startups from zero to a production-grade platform — secure AWS landing zones, Kubernetes that scales, reliability you can sleep on, and AI-native tooling baked into the workflow. Senior, hands-on, owning the problem end to end.
terraform apply -auto-approve
Provisioning acme-startup landing zone…
VPC + multi-AZ networking
EKS control plane · Karpenter autoscaling
Istio mesh · mTLS + rate limits
RDS · ElastiCache · IAM baseline
Observability + RBAC (Datadog)
CI/CD + GitOps pipelines
Apply complete. 42 added, 0 destroyed.
→ platform ready for first deploy
What I do
Infrastructure as a service offering, not a slide deck.
Four ways we plug into a startup — usually all at once. Hands on the keyboard, decisions documented, no hand-off to a team that doesn't exist yet.
Kubernetes platforms
Production EKS that scales sanely — Karpenter node provisioning, Istio service mesh, GitOps delivery, and developer-friendly paths to ship. Built so your engineers move fast without owning the cluster internals.
Reliability & SRE
Observability that means something, alerting that doesn't cry wolf, and deep production debugging across the stack — CNI, mesh, queues, databases. I treat failures as fixable root causes, not things to architect around.
AI-native DevOps
Bringing AI into the operational loop where it actually pays off: a Slack-native assistant that queries your clusters, observability and Git in real time; AI-assisted developer tooling; and rigorous evaluation (LLM-as-judge, RAG eval) so model-driven workflows stay reliable instead of hand-wavy.
Where we go deep
Depth where startups get stuck.
The failure modes that don't have a quick forum answer — networking, mesh, autoscaling, state, access. This is the level we operate at, so your team doesn't have to.
Kubernetes networking & scale
When ENI limits, IP exhaustion, or autoscaling break in ways no forum thread covers, that's home turf. We work the real cause — security groups, VPC endpoints, IAM, node lifecycle — instead of bolting on a workaround that bites later.
Clusters that scale predictably, not surprises at 2amSystems that stay up
Service mesh, rate limiting and health checks designed so the platform degrades gracefully when a dependency doesn't. Resilience built into the architecture, so an outage downstream doesn't have to become an outage for you.
Reliability you can actually sleep onSecure access, no VPN sprawl
Get your team into Kubernetes and databases safely, without legacy VPN overhead — modern access patterns wrapped in tooling engineers actually want to use. Tight for security, frictionless for developers.
Least-privilege access that doesn't slow anyone downAI in the operational loop
AI tooling grounded in live system state — clusters, observability, code — and rigorously evaluated so model-driven workflows stay trustworthy. The genuinely useful end of AI for DevOps, not the hype.
Faster operations, without the guessworkGovernance as code
Access, permissions and observability codified and version-controlled, so control scales with the organisation instead of drowning in tickets. The guardrails are in the repo, not in someone's head.
Control that grows with you
How I work
Methodical, incremental, honest about trade-offs.
/ 01
Find the constraint
Take the ambiguous problem and reduce it to the one thing that actually limits you. No solution before the real question is clear.
/ 02
Design for reality
Pragmatic architecture grounded in best practice — built for the team and budget you have now, not a hyperscaler you're not yet.
/ 03
Ship incrementally
Small, safe, reversible changes. Feature flags and staged rollouts over big-bang launches. Verified in production, not just in theory.
/ 04
Leave it documented
Concise runbooks and clear decisions, so the platform stays operable long after the engagement ends. Code that's safe to evolve.
Tooling & environment
The stack I bring.
Cloud & platform
- AWS · EKS · VPC CNI
- Karpenter · NodePools
- Istio service mesh
- Terraform · Atlantis
- Helm · GitOps · CI/CD
- Bazel builds
Reliability & data
- Datadog · observability
- Production debugging
- RDS / PostgreSQL · WAL/CDC
- ElastiCache Redis
- Kafka Streams · Connect
- NVMe vs network storage
Code & AI
- Go · Python · TypeScript
- gRPC · protobuf · React
- CLI & internal tooling
- MCP servers / clients
- LLM eval · LLM-as-judge
- RAG eval · AI-assisted tooling
About
A boutique with senior depth.
We treat failures as fixable root causes — and bring a designer's eye to the engineering, not just the architecture diagram.
Digital Cloud Consulting is a focused platform-engineering practice for startups. We work across the full surface — from low-level cloud networking and Kubernetes internals to developer tooling, observability, and AI-assisted operations — with the seniority to make the calls that matter early.
The difference is depth and ownership. You get engineering that takes an ambiguous problem, finds the real constraint, and carries it through to production — no hand-off to a team that doesn't exist yet, no workaround left for future-you to untangle.
Methodical, incremental, and honest about trade-offs. Founded and led by Dario Tišlar.
Good fit when you need to
…go from greenfield to grown-up.
- Stand up a real AWS foundation before you scale onto sand
- Get Kubernetes that your team can ship on without firefighting
- Turn flaky production into something you can actually trust
- Add AI to DevOps workflows without the hype or the breakage
- Borrow a senior platform engineer without the full-time hire
Let's build
Got a platform to get right?
Tell me what you're building and where it hurts. I'll come back with a concrete, no-fluff take on the path from here.
dario@digitalcc.hr →or reach out on LinkedIn