Sarim Zafar Consulting
Your cloud and GPU platform,
made boring.
I am a senior SRE and AI-infrastructure engineer. I cut cloud and GPU spend and make shaky platforms boring and reliable, so your team gets back to shipping. No new hire needed. 12+ years on AWS, Kubernetes, and GPU/LLM systems.
Trusted by teams at
When you grow faster than your platform, it costs you.
It shows up as cloud bills you cannot explain, GPU spend that leaks, and 3am pages nobody signed up for. Releases slow down. And hiring a senior SRE to fix it takes six months you do not have.
What I do
Cloud Solutions & Migration
Architecture and migrations that hold up under real load, on AWS, GCP, and Kubernetes.
DevOps & LLMOps Engineering
GPU/LLM platforms, CI/CD, observability, and serious cost control at scale.
Custom AI Agent Development
MCP, RAG, and agents wired into your real systems instead of throwaway demos.
AI Audit & Transparency
Know what your AI does, why it does it, and where the risk sits.
Results
A few numbers from past projects.
served on the Scout24 platform, about 110K per hour against GPU nodes.
cloud & spot spend cut at Scout24 (800+ K8s clusters, GPU/LLM platform).
in-store theft reduction across 27,000 Kaufland & Lidl stores.
SLA maintained across BMW platform operations.
customers in 2 years at Paessler PRTG (hosted SaaS).
vehicle sensors at scale with cross-region DR at Elektrobit.
high-performance computing research at TU Dresden.
How we work together
Most people start with a Readiness Review. It tells you what's worth fixing before you commit to anything bigger.
Start here
Cloud & AI Readiness Review
A fixed-scope review of your AWS, GCP, and Kubernetes setup, done in about two weeks.
- ✓ Where your cloud and GPU spend is leaking
- ✓ Whether your platform and AI are built to scale
- ✓ A short report and a 90-day plan
€2,900
Credited toward your next engagement.
Build and fix
I implement the roadmap with you, hands-on.
- ✓ Cost and reliability fixes shipped
- ✓ Larger work as fixed-fee projects
- ✓ Or a share of what I save you
Fixed fee or % saved
You pick what fits the work.
Fractional platform engineer
Ongoing senior platform help, part-time and embedded in your team.
- ✓ Cover the gap while you hire
- ✓ Cost, reliability, and GPU/LLM ops
- ✓ Flexible, month to month
From €5,000 / month
Billed monthly.
What people say
"As talented at creating elegant solutions to complex problems as he is dedicated to supporting his team. He onboarded faster than all expectations and, unprompted, documented what he learned so the whole team benefited. An asset to any team, and a natural born leader."
VP Product Development, Paessler AG
"Sarim's expertise was pivotal in developing a future-proof architecture for our application. He thinks ahead, he is technically strong, and he handles complex work on his own. I highly recommend him."
Product Owner, Fielmann AG
"An exceptional talent, making both technical and cultural contributions to our team. DevOps is a culture of collaboration, and Sarim embodies those values."
VP Software Engineering, Bellwether Coffee
"Sarim delivered excellent results and stayed persistent and hardworking, even on the tedious tasks. When we had important milestones to hit, I could always rely on him. Eager to learn, and focused on the outcome."
Project Manager, Elektrobit GmbH
"Sarim is an optimistic, kind, and highly skilled professional. Working with him is fun and goal oriented. He was a great addition to our development team."
Client, Accenture engagement
"A sincere and dedicated developer with real character. A self-starter and a budding leader who needs very little guidance. A great addition to any team."
Software Engineer, Elektrobit (RVD project)
"Sarim is an outstanding developer and truly a pleasure to work with. He is responsive, thoughtful and creative, and suggests new approaches to solve the problem. Trustworthy and accountable. I highly recommend him."
Logical LLC, Ohio
"Sarim took ownership of the tasks assigned to him and ensured we completed the project on time."
CTO, Luzid.Tech GmbH
About
I am currently Senior SRE at Scout24, running GPU and LLM platform infrastructure. Over the past 12 years I have worked with BMW, the Schwarz Group (Kaufland and Lidl), Fielmann, RWE, Paessler and Elektrobit.
I spent around 15 years living and working in Germany and hold an M.Eng from a German university. I now run my consultancy from Bucharest, where I live with my family. I work in EU time zones, travel to DACH when it helps, and speak fluent German. Google Cloud certified, and an iSAQB Certified Professional for Software Architecture. I am also a published IEEE author.
FAQ
What does Sarim Zafar do?
I am a senior SRE and AI-infrastructure consultant. I help scaling companies cut cloud and GPU spend, make platforms reliable, and ship faster, across AWS, Kubernetes, and GPU/LLM systems.
Where are you based?
I am a German national, now based in Bucharest. I work in EU time zones and travel to DACH when it helps.
What results have you delivered?
The Scout24 platform I run serves about 1 billion requests a day, with roughly 110K an hour hitting GPU nodes. I cut cloud and spot spend by about 40% there, helped reduce in-store theft by 30% across 27,000 Kaufland and Lidl stores, kept a 100% SLA at BMW, and helped scale Paessler PRTG from 30 to 16,000 customers in two years.
What do you offer?
Cloud solutions and migration, DevOps and LLMOps engineering, custom AI agent development, and AI audit and transparency.
How do we start?
Book a 30-minute intro call, or email contact@sarimzafar.com. Tell me where your platform hurts and I will tell you straight if I can help.
Worth a quick call?
Tell me where your platform hurts. I will tell you straight if I can help.
Not ready to talk? Subscribe to my newsletter, Cloud Catalyst.