Most cost work ends with a report. Someone else still has to do the work. I do both parts: I find where the money goes, then I go in and change the infrastructure and the .NET code that spends it.
A dashboard, a tagging policy nobody enforces, and a list of rightsizing recommendations. The recommendations are usually correct. They land on a team that already has a backlog, so six months later the document is stale and the bill is bigger.
I write the Terraform, the policies and the pipelines myself. If a service needs a bigger VM because it leaks memory, I open the solution and fix the leak. You end up with the saving, not the plan for one.
Example: debugging an undocumented AzureRM bug in a Terraform migration →01 · Start here
Two weeks, fixed price. I go through your subscriptions and tell you where the money actually goes and what's worth fixing first.
02 · Implementation
Everything from the assessment, carried out as code in your repositories and your pipelines.
03 · Application level
Some spend can't be fixed from the portal. When the architecture is what costs the money, the fix is in the code.
Rebuilt foreign-currency transaction processing for a bank as a distributed REST API. It freed up about 0.75 FTE of operational work every year.
Cut an airline application's memory use by over 70%, which made a smaller and cheaper instance class viable in production.
Moved legacy microservices onto right-sized, elastically scaled infrastructure. Less operational overhead, smaller compute footprint.
Thirty minutes about your estate, your bill and what's worrying you. No slides.
Reader rights on the subscriptions in scope. I can't change anything, and at this stage I don't need to.
Two weeks to findings, sized savings and a plan. If you want to take it from there yourself, that's a fine outcome.
Optional. I do the work alongside your engineers and hand over the code and a runbook.
Send me last month’s Azure bill and one line about what worries you. You’ll get an honest answer about whether there’s anything worth doing here, including when there isn’t.
Based in Novi Sad, Serbia. Working remotely across CET and US hours.