Keeping up with CVEs in production
Smaller images, faster fixes
Start with hardened images and scan them before deployment and while they run in production. Test patches in CI so you can ship fixes quickly.
Read the articleSoftware engineering · Production systems
Beyond the happy path.Lessons from production
I build high-performance backends and distributed systems. I helped develop and run a national eHealth platform that serves 50 million users. We had to meet strict performance SLAs and security requirements. I write about engineering decisions in production and how AI changes the way we make software.
Read my articles// production has more than one path
request
├── success
│ └── persist result → release resources
├── timeout
│ └── retry safely → backoff + jitter
├── overload
│ └── bound concurrency → apply backpressure
├── dependency failure
│ └── fail fast → recover when healthy
├── shutdown
│ └── stop accepting → drain in-flight work
└── restart
└── reclaim unfinished work → resume safely
On my mind lately
All articlesHigh-performance backends in the age of AI
10 principles that still matter
When requests start piling up, I look at where they spend their time and how much work the service takes on at once. The same checks apply when AI wrote the code.
Read the articleWhat I check before shipping code written with AI
18 questions about production reliability
What happens if the service restarts halfway through a request? Questions like this help me decide what to ask AI to build and what to test before release.
Read the article
About me
I’m Razvan, a software engineering lead and software architect with more than 11 years of experience building and operating backend systems in financial services and healthcare.
I work across architecture, implementation, and production operations. That can mean tracing a slow request from the service to the database, investigating why a job ran twice, or reviewing how a release will behave under load or when a dependency fails. I also review code written with AI coding agents and help teams decide what to verify before it ships.
I’ve also mentored more than 200 interns and continue to coach engineers across teams.
Problems I help teams solve
Backend performance
A slow PostgreSQL query, rising response times, or a service that struggles as traffic grows. I can help find the bottleneck and work out how to measure whether a fix made a difference.
Production reliability
Recurring incidents, unreliable background jobs, or failures that can’t be reproduced locally. I can help trace what happened and identify what would make the system easier to operate and recover.
Architecture and release reviews
A design decision with competing tradeoffs, a risky rollout, or code that needs a second look before it reaches production. I can help you assess the failure paths and decide what to test before you ship.
Security hardening
I can help harden backend services and cloud configurations, prioritize and remediate CVEs in dependencies and container images, and apply zero-trust principles with verified service identities and least-privilege access. I can also review how your system handles secrets, enforces authorization, and keeps tenant data isolated.
Tell me what you’re working on
If one of these sounds familiar, email me or send me a message on LinkedIn. A few lines about what you’re building, where you’re stuck, and what you’ve tried are enough to start.