INFRASTRUCTURE AS CODE, NOT ARCHAEOLOGY DATA CENTER TO CLOUD WITHOUT THE DOWNTIME
Your ops team spends more time fighting infrastructure than building on it. Deployments are manual rituals. Your cloud bill is a mystery. And that 'disaster recovery plan'? It's a wiki page from 2021 that starts with 'Step 1: Call Dave.'
You need someone who's migrated systems that can't go down, and knows the difference between cloud-native and cloud-wasteful.
Infrastructure Signals That You're Fighting the Wrong Battles
I've tracked these patterns through startups and enterprises alike. If more than two sound familiar, your infrastructure is working against you.
Manual Deployments
Deploying to production requires a specific person, a specific sequence, and a specific prayer. Your deployment runbook has 47 steps, half of them are 'if this fails, try...' and nobody's sure which version is current.
Config Drift
Your staging environment doesn't match production. Your production servers don't match each other. Someone SSH'd in six months ago to 'fix something quickly' and now that fix is load-bearing infrastructure nobody documented.
Cloud Bill Surprises
Every month, finance asks why the AWS bill went up 30%. Nobody knows. Orphaned resources, oversized instances, and that dev cluster someone spun up 'for testing' eight months ago are all still running.
No Disaster Recovery
If your primary region goes down, your recovery plan is 'hope it comes back.' No tested failover, no automated backups verification, no runbooks that anyone has actually rehearsed.
From Manual Archaeology to Infrastructure as Code
AWS, GCP, Kubernetes, Terraform
I don't just move your servers to the cloud and call it a migration. I redesign your infrastructure for the patterns that actually matter: immutable deployments, auto-scaling that responds to real metrics, observability that catches problems before your users do.
From bare-metal migrations to Kubernetes orchestration, Terraform modules to CI/CD pipelines that your team actually trusts — I've tracked these trails through banks, fintechs, and scale-ups where downtime means lost revenue.
When I leave, your infrastructure runs itself — and your team knows how to evolve it.
- Charting the Cloud Trail: Lift-and-shift or re-architecture: the right strategy for your actual constraints, not the vendor's sales pitch
- Infrastructure as Code: Terraform, Pulumi, CloudFormation. Version-controlled, reproducible, reviewable infrastructure
- Container Orchestration: Kubernetes, ECS, managed services. Right-sized for your team's capacity to operate
- Tracking What You're Running: Monitoring, alerting, cost tagging, right-sizing. Know what you're running and what it costs
The Migration Trail
Cloud migration isn't a lift-and-shift weekend project. Here's the trail we follow: five phases that take you from where you are to where you need to be.
Assess Current State
- Full infrastructure audit: what's running, where, and why
- Dependency mapping across services and data stores
- Cost analysis: current spend vs. projected cloud costs
- Risk assessment: what breaks if we move it?
Design Target Architecture
- Cloud-native patterns sized for your team's capacity
- Network topology, security model, access controls
- Multi-environment strategy: dev, staging, production parity
- DR and failover design: tested, not theoretical
Build IaC Foundation
- Terraform modules for every infrastructure component
- CI/CD for infrastructure — PRs, reviews, automated applies
- Environment parity: staging mirrors production exactly
- Secrets management, encryption, compliance controls
Migrate & Validate
- Priority-ordered workload migration: high-value first
- Data migration with validation and rollback plans
- Load testing in the target environment before cutover
- Zero-downtime traffic shifting with instant rollback
Optimize & Transfer
- Cost optimization: right-sizing, reserved capacity, spot usage
- Auto-scaling tuned to real traffic patterns
- DR testing: actually run the failover, don't just document it
- Team training: your engineers own this infrastructure
Picture This
Your deployments are one-click and reversible. Infrastructure changes go through pull requests. Your cloud bill makes sense because every resource is tagged, tracked, and right-sized.
Disaster recovery is tested quarterly — not theoretically planned. Your ops team builds platforms instead of fighting fires. New environments spin up in minutes, not days.
Companies That Grew Through The Breaking Point
Real companies, real pressure, real results. Here's what happened when they brought me in during their scaling challenges.
10x Scale During Hypergrowth
The company was growing fast—too fast for their architecture. I led the redesign of core systems, migrated to managed cloud services, and introduced TDD practices. Four-year engagement that took them from scrappy startup to enterprise-grade platform.
Internal AI Assistant for 50,000+ Employees
Architected enterprise productivity tools on AWS with full Terraform IaC. Built high-performance backend with comprehensive load testing frameworks to handle bank-wide rollout across all regions.
"Szymon built the reference architecture templates that twenty other teams ended up adopting. He didn't just ship his own service — he turned what he built into the golden path for the entire programme. His team went from middle-of-the-pack to the highest-performing unit in the delivery." — Delivery Lead, Global Travel Technology Provider
Why This Works When the Migration Trail Gets Lost
Why Not Cloud Vendors?
- They sell services, not solutions. Their architects recommend their products, their training covers their tools, and their success metrics measure your consumption, not your outcomes.
- I'm vendor-agnostic. I recommend what works for your constraints, not what earns me a referral bonus.
Why Not Internal Ops?
- Your ops team is too busy keeping the lights on to redesign the electrical system. They know the problems — they just don't have the bandwidth to fix them properly.
- I bring the bandwidth and the battle scars. Your ops team gets a partner, not another set of demands.
Let's Chart the Migration Trail
Your migration won't get simpler next quarter. 30 minutes. Let's chart the trail from where you are to where your infrastructure needs to be. No vendor pitch, no pressure.
Prefer email? contact@grupavii.eu