$360K / year (2025)
Infrastructure waste nobody was looking for. Self-identified. Now savings are built into the platform.
€20K / day (2017)
Cash leaking daily. Internal tooling couldn't find where or how much.
20x less downtime (2025)
100% uptime during Super Bowl.
10,000 man-days / year (2021)
Multi-year rewrite had failed. Built inside existing system. Still running.
Andrei Ursuleanu
What's broken in your platform that nobody has questioned yet?
↓
The problems nobody owns
Your platform has problems nobody thought to question. They sit between systems, between teams. Unclear, expensive, and compounding every month.
I debug what is really happening in the service or platform. Then I make the real fix the new default — not a patch someone has to keep alive.
I've done architecture, performance, devops, testing, sales, support, contracts, hiring, even litigation. Not at the same company. But the person who's done all of them sees problems differently than someone who's done one.
Problems solved
Infrastructure costs growing and nobody knows why?
At Adobe, nobody had traced where the waste was actually happening. I debugged the running Prometheus setup across tenants, found the waste in the default allocation, and changed the model. Savings are recurring. The fix became the foundation for ScaleOps integration. New tenants inherit it by default. Anyone who wants to bypass it has to explicitly opt out.
$360K / year saved (2025)
High-risk migration across hundreds of services?
At Adobe, PagerDuty→ServiceNow migration looked routine to everyone else. I flagged the risk: a single missed line would lose alerts on production services. Then I collaborated with 150 teams to create a unified approach. Replaced hundreds-of-changes-per-tenant with a single-property configuration. For 150 teams, it's a one-liner. Now the standard.
150 teams, zero missed alerts (2025)
Money leaking and your team can't find where?
A Romanian land-based gaming operator knew money was leaking. Not how much. Not where. Internal tooling didn't catch it. I scoped the full path from prize verification to payout to cash machine reconciliation. Built observability that made the leak visible for the first time. Built both business and technical processes. Platform runs without me.
€20,000 / day in cash losses — now they knew exactly where (2017)
Legacy system everyone's afraid to touch?
At Dennemeyer, a multi-year rewrite had failed. I built the automation against constraints within the existing monolith. People freed to handle new IP types — patents, designs. Built inside the existing system, so nobody needed to learn anything new. Still running.
10,000 man-days / year saved (2021)
Performance degrading as you scale?
At Eved, I owned platform performance serving Apple, Dell, PwC. Data volume grew 5x during my tenure. The team became self-sustaining — they know how to measure, what to apply. No dedicated performance person needed anymore.
10x improvement on critical paths (2017)
Clients not renewing because delivery keeps missing what they actually need?
At 3 Screen Solutions, sales sold and devs built — but nobody synced directly with the client's technical teams. Constant rework, slow delivery, low renewal rates. I started talking to their engineers directly — even doing uncontracted work to close the gap. Friction dropped, delivery speed went up, clients renewed with bigger contracts. Swisscom, Siemens, Kabel Deutschland. Platform core survived — after many iterations, now powers Google TV OS.
25% of future revenue (2013)
What I build
The platform underneath the product. Shared cores that multiple teams ship on top of. Observability that tells you what's actually happening — not what you hope is happening. Performance baselines that hold when data grows 5x. Deployment pipelines that make the risky thing safe. Cost controls that run by default, not by request.
I've built these across backend services, cloud-native infrastructure, distributed systems, and embedded devices. The layer changes. The discipline doesn't.
How I work
After 20 years, I noticed the same shape in every project that worked. I call it QSAIL — a five-step method for solving problems that are unclear, misunderstood, or being solved wrong.
Q — Question
Challenge the premise before touching anything. "Why are we solving this at all?"
S — Scope
Map who knows what. Build the full picture before solving.
A — Architect
Design the solution AND the guardrails. Architecture and process as one system.
I — Implant
Embed in people and systems so it survives without me.
L — Lift
Leave everything stronger than I found it.
Not a consulting framework. It's what actually happens when I work.
Explore QSAIL
Your problem has been solved before
A scaling problem on a set-top box, a data integrity problem in IP management, and a cost problem in cloud infrastructure can have the same shape.
Consumer & Media
Smart TV, automotive retail, IoT, micro-mobility
Industrial & Precision
Scientific instruments, semiconductor robotics
Enterprise B2B
SaaS platforms, IP management, experience platforms
Heavily Regulated
Healthcare, fintech, regulated gaming operations
When a CTO says "we've never seen this problem" — I sometimes have, in a different industry. The pattern is the same. The label is different.
Education
The education formalized what 20 years of practice had already taught.
🇺🇸
MIT Sloan
Management, Innovation, and Technology (2023–2025)
🇨🇭
IMD
Driving Strategic Innovation (2025)
UNITBV
MBA (2009–2011)
UNITBV
BSc Electronics & Computer Science (2004–2009)
What people say
"Considers the bigger picture while tackling individual tasks. Particularly effective when given ownership of larger project modules."
— VP of Product Engineering, fintech platform
"When I knew he was in charge of a review or coding a requirement I knew things will be done right."
— Product Owner, IP management firm
"Among the first persons I could consider a mentor. Technical knowledge proven from the start with clean and innovative solutions."
— Developer, scientific instruments company
"If I would choose a word to describe Andrei, it would be: Critical Thinking"
— most of my past colleagues
JoyDev
I founded JoyDev — the lab behind some of my product and advisory work.
My own money. My own risk. Some ideas shipped. Some didn't. All taught something real.
See what JoyDev builds
Let's talk
If there is a service or platform problem nobody can pin down, I can help debug it and make the fix stick. I work as advisor, architect, or hands-on — depending on what the problem needs.
I engage through retainers, project-based work, or equity partnerships. First conversation is always free.