Your AWS bill goes up every month? Get the money back.
First I cut your AWS costs at a fixed price. Then I rebuild your application so load peaks, new customers and new releases stop being a risk.
References: Large Scale E-Commerce, Builders & Creatives
- 5 min
- instead of 24 hours to fresh datacontinuous processing instead of a nightly run
- -45%
- AWS billright-sizing, Spot and automatic shutdown
- 99.95%
- availabilitymeasured over twelve months
- 0
- failed overnight runssince the switch
Does this sound familiar?
For companies running their own software on AWS, such as a customer portal, an internal platform or a SaaS product, that has become too expensive or too slow.
- The bill grows faster than the business
- A few percent more every month, and nobody knows exactly why. Servers run at night and at weekends when nobody needs them.
- Releases are a risk
- New versions go live by hand, preferably on a Friday after hours. Every bug means stress.
- Load peaks bring everything down
- One campaign, a month-end close or one more large customer, and the application slows down or falls over.
- A new customer means a new installation
- Your SaaS product runs on separate infrastructure for each customer. Every new customer makes operations more expensive and more complicated.
What changes for you
- Your application stays fast, even at month end and in the middle of a campaign.
- New versions go live on Tuesday at 10 a.m. instead of late on a Friday night, and a faulty release is rolled back in minutes.
- You know what each customer costs in infrastructure and can set your prices accordingly.
- The cost optimisation usually pays for itself within a few months.
Where I have done this before
Large Scale E-Commerce
24 h → 5 min until new product data is live
Moved a product data pipeline for millions of products from a nightly batch to real time. No more failed overnight runs, and data from today instead of yesterday.
Read the case studyBuilders & Creatives
-45% GPU cost after right-sizing
Took MLOps pipelines and AI models into production on AWS and ran them there, on a platform I designed myself.
Tomislav built and operated the cloud infrastructure our platform runs on, solving real scaling challenges along the way. A dependable architect we could always count on.
Sascha HalvaCEO Halva InvestHow the modernisation runs
I work in stages that each deliver value on their own. Your application stays live throughout, and your development team works with me from day one so the knowledge stays in house.
Automate
Infrastructure as code and automated delivery. From here on, every change goes live tested and repeatable.
Decouple
The parts that slow everything down are split out and scale independently of each other.
Go multi-tenant
For SaaS products: one platform for all customers, cleanly separated data, measurable cost per customer.
Observe
Monitoring and alerts, so problems show up before your customers call.
What I need from you
- At least one person from your development team working with me
- Access to the repository, AWS account and ticket system
- One weekly check-in
Technical scope for your IT team
- Target architecture and architecture decision records
- Infrastructure fully in Terraform or CDK
- CI/CD pipeline with automated tests
- Containers on ECS or EKS, serverless where load fluctuates
- Monitoring, logging and tracing
- Tenant isolation and cost per tenant (SaaS)
- Load tests with documented limits
- Handover to your development team
Start small, then decide
Every step delivers its own result at a fixed price. You can start or stop at any point.
AWS cost check
A questionnaire about your bill and your environment, with the estimated savings in euros.
- Savings potential per month and per year
- Your biggest levers in euros
- No access to your systems needed
Fixed-price cost optimisation
I put the quick wins in place myself instead of just sending you a list.
- Savings already implemented
- Cost dashboard and budget alerts
- Usually pays for itself within a few months
Architecture review
A review against the AWS Well-Architected Framework, focused on scaling, resilience, releases and cost.
- Where your application hits its limits today
- Target architecture as a diagram
- Rebuild plan in stages, with effort estimates
Scalable cloud software
Modernise your existing application or build a new multi-tenant SaaS platform.
- Handles load peaks without manual work
- One-click releases with a way back
- Cost per customer you can measure
All prices exclude VAT. "From" prices depend on scope and become a fixed price after the assessment.
You work directly with me
I'm Tomislav Pree, a cloud architect from Mainz. I have built and run cloud platforms since 2018, most recently for Builders & Creatives, in large scale e-commerce and in the aviation industry. From the first call to the handover you talk to the person who does the work.
Afterwards: Part-time cloud architect
For teams that have developers but no experienced cloud architect. Two to four days a month, booked in advance.
- Architecture decisions and reviews
- Code and infrastructure reviews
- Monthly cost control
- A sparring partner for your developers
What clients ask before they start
Do we have to rewrite our application from scratch?
Almost never. In most cases it is enough to automate the infrastructure and rebuild the specific parts that slow things down. A rewrite only pays off when the old base will not carry you through the next few years.
How much can we save on AWS costs?
In environments that have never been optimised, 20 to 30 percent is common. The free cost check gives you a first estimate of what that means for you.
Will you work with our development team?
Yes. I work in your repository, with your tools and within your processes. The aim is for your team to run the platform themselves afterwards.
How much of our time will this take?
You need one fixed point of contact who can make decisions or get them made quickly. At the start and at handover that means a few hours a week, in between usually one short meeting a week. I do the actual work.
Kubernetes or serverless?
Your load and your team decide that, not the latest trend. For many business applications, ECS with Fargate is the pragmatic choice. Kubernetes pays off beyond a certain size, serverless when load swings sharply.
Further reading
- GitOps with ArgoCD and Helm: Auditable Releases for Regulated IndustriesGitOps with ArgoCD and Helm makes every release auditable: who approved what and when is in the commit. Rollback through Git, security scans in the CI pipeline.
- Setting Up Kafka Right: Topic Design, Partitioning, and IdempotencyKafka topic design decides your data quality: how partition keys, schema rules, and idempotent consumers prevent duplicates and hot partitions in production.
- Reduce AWS Costs: Right-Sizing and Spot Instances Done RightReduce AWS costs without risk: right-sizing based on real utilization data, spot instances where they are safe, and budgets that flag runaway spend early.
Where do you stand today?
Enter your monthly AWS costs, answer eight questions and see your biggest savings levers in euros straight away.