
- running production cloud on AWS, Azure, GCP and Oracle Cloud
- 8 years
- fewer unplanned outages at Niche Academy
- 90%
- traffic headroom found with load testing at IDWise
- 3x
- uptime SLA held on an identity verification platform
- 99.95%
Where I do my best work
I take on fractional CTO and architecture work, but reliability is where I add the most. These are the four things I improve.
Performance
Load tests that model real traffic, p99 latency and throughput work, and capacity planning done before the launch rather than during it.
Availability and resilience
SLOs that match what customers feel, observability you can debug with, and failover that has been tested before you need it.
Developer experience
Terraform, CI/CD and Kubernetes platforms that engineers trust enough to ship to every day, so releases stay small and safe.
Cloud cost
Rightsizing, scheduling and architecture changes that lower the bill without trading away any of the above.
Ways to work together
No account managers and no handoffs. The person who scopes the work writes the Terraform, runs the load tests and presents the findings.
Book a free 30-minute callFractional CTO
Part-time senior leadership. Architecture decisions, the infrastructure roadmap, vendor choices, hiring, and a technical voice with investors and customers.
Solutions architecture
A review of your AWS, Azure or GCP setup against reliability, performance, security and cost, with written findings and a plan your team can act on.
Reliability project
A defined outcome. A launch that has to hold load, an observability rebuild, a Kubernetes migration, or a multi-cloud deployment for compliance.
Industries and teams
Eight years across startups, scale-ups and large enterprises. The experience page has the details for each one.
Industries
Teams I have worked with
Founder of upzero.io and md0.io
Building my own products keeps my advice honest. I pay the cloud bill, answer the alerts and make the same trade-offs I help clients make.
Monitoring and observability for infrastructure and AI pipelines. Uptime, latency and incidents checked from five regions, with status pages that update themselves.
md0.ioLiveFree Markdown tools that run in the browser, plus a GitHub-native CMS that edits the files in your repo. No server round trip, no tracking.
Latest writing
All postsDoes your product slow down or go down at peak hours?
That is the problem I most like to solve. Tell me what happens when traffic is highest, and in 30 minutes we can work out where to look first.
Safoor Safdar




