Draft · not final

AWS cost optimization by an ex-AWS engineer

My name is Cristian Magherusan-Stanciu and I specialize in optimizing large-scale AWS setups.

I do the work myself, using tools I built over 12+ years working on cost reduction - first for major tech companies, then for AWS itself.

I never outsource. And I never leave any work for you to do yourself.

I find the safest, highest value changes and implement them myself (through configuration or tiny PRs that take 5 minutes to review).

Drop AWS cost this month, not "when there's time in the sprint"

See my 49-point cost saving checklist

Cristian Magherusan-Stanciu

2020 – 2022

Sr. Specialist Solutions Architect (Spot & Graviton), AWS

Helped AWS's largest customers cut compute spend with Spot and Graviton.

2019 – 2020

Sr. DevOps Consultant, AWS Professional Services

Migrated security-sensitive financial-industry customers to AWS.

2010 – 2019

Principal Engineer, HERE / Nokia

Migrated large-scale location services to AWS and optimized for cost and performance.

Open source projects

AutoSpotting Replaces on-demand EC2 with Spot. Over $100 million saved in aggregate.
ec2instances.info Top reference site for EC2 instance pricing and specs (in collaboration with Vantage).

LinkedIn →

How this works

What you get (and don't get)

No fancy dashboards (they're nice to look at, but who'll do the work?)
No AI touching production environments (because who has time to recover from a random rm -rf?)
No meetings, updates, reports, or standups - fully done-for-you
No budget approval needed (paid through AWS Marketplace out of savings)
No outsourcing, offshoring, or cheap Fiverr "specialists"

Just one ex-AWS engineer, finding the savings and shipping them himself.

Recent wins

What this looks like in practice

A US fintech scale-up ($3M/year on AWS)

Found a production database still running for a B2B customer who had cancelled a year earlier, and shut it down. Saved $378k/year.
Restructured Reserved Instance coverage across RDS. Saved $161k/year.
Moved a steady-load database off Aurora Serverless back to provisioned capacity. Saved $125k/year.

Avoided $970k/year in total - about a third of their AWS bill.

A US e-commerce SaaS scale-up

Found they were overpaying for Aurora I/O-Optimized, and rolled it back to only the databases that benefited. Reduced the database bill 30%.
Right-sized ElastiCache and moved to Valkey. Reduced cache costs 60%.
Moved S3 data to the right storage tier. Reduced storage costs 63%.

Reduced the bill $178k/year - about 30% overall.

A gig-economy SaaS (Series A)

Cleared years of stale OpenSearch data and right-sized the cluster. Reduced their single biggest cost by half.
Right-sized ElastiCache and moved to Graviton. Reduced cache costs 90%.

Reduced the bill $150k/year - about 45% overall.

A GenAI startup (seed stage)

Consolidated standalone EC2, 70% of the spend, onto ECS and Spot with savings plans.
Right-sized RDS, moved to Graviton, and migrated IO1 storage to GP3.

Reduced the bill from $12k/month to under $4k - two-thirds.

A marketing-compliance SaaS ($1M/year on AWS)

Fixed CloudWatch Logs retention, same day. Reduced it 90%.
Cleaned up S3 and moved data to the right storage class. Identified 50% off the biggest driver.

Identified $300k/year in savings.

Do it yourself

Get the 49-point checklist I run against every account I review, and go through yours line by line.

Get the 49-point checklist

See what you could save

Send me an anonymized Cost Explorer export and I'll tell you the three biggest things I'd fix.

Upload Cost Explorer CSV

Common questions

Do I have to give you access to my AWS account?

Not to start. Send an anonymized Cost Explorer export and I'll tell you the three biggest things I'd fix - before you give me anything. Access only comes later, if you decide to have me do the work.

Will you touch production? Could you break something?

I work config-first: the safe, reversible changes first, biggest and lowest-risk before anything else. Anything that needs code goes as a small PR you review in five minutes. No automated tools let loose on prod, and no big code changes.

Isn't this just another consultant who drops a report and leaves?

No. I do the work and implement the changes myself - you don't get a report with a to-do list at the bottom. No juniors, no handoff, no outsourcing.

Why not just use AWS's own tools, or AI?

They tell you what you already half-know and leave the doing to you. I don't trust AI touching prod, do you? The savings come from judgment - what to keep, what to kill, what's safe to change - not from a dashboard.

How do you charge?

A percentage of what I actually save you, billed through the AWS Marketplace. No upfront cost, nothing to approve. I only make money when your bill goes down.

I could do this myself. Why would I hire you?

You can - the checklist above is exactly what I'd run. The question was never whether you can. It's whether you have time to do it - this sprint, the next one, or the one after that.

Ignorance is bliss

Not knowing works, for a while. As long as you don't know how much you're overpaying AWS, you don't have to do anything about it. Once you know, it's over. You can't unsee paying $20k, $50k, $100k a month more than you have to, and you can't unsee that it was there the whole time.

So there are three options:

Don't look. Just keep paying until money runs out (or someone else notices).
Look, and fix it yourself. The checklist above is the whole list. You can do it, one item at a time.
Look, and let me fix it for you. Send me the export and I'll tell you the three biggest things I'd fix.

Upload your Cost Explorer CSV