About Me

In 2018, I built a system that could predict when a vehicle's components would fail. Not from a lab, but from sensors running under real conditions with real consequences for getting it wrong.

That project taught me something that's shaped everything since: sophisticated algorithms mean nothing if the hardware fails, the data pipeline breaks, or the operator doesn't trust the system.

I founded Tigon to build infrastructure that earns trust through consistent performance rather than impressive demos. Today, I'm building sensor-based monitoring systems that run autonomously at the edge.

My work started with a university thesis on sensor-based component monitoring and failure prediction. The premise was straightforward: if you could continuously observe physical systems, you could predict and prevent failures before they happened.

Building that vehicle monitoring platform taught me that prediction isn't the hard part. The hard part is measurement infrastructure, data quality, inference under uncertainty, and operational integration. It's making systems that work after you've left the room.

Most infrastructure fails not from bad algorithms but from overlooked realities: power fluctuations, network dropouts, sensors that drift, operators who have 50 other things to monitor. I build for those conditions.

I think about hardware, software, and operations as one system. A dashboard that looks beautiful but crashes when the database hiccups isn't a product, it's a demo. A sensor that works in the lab but fails in humidity isn't ready.

I prefer boring technology that compounds over years to exciting features that ship in weeks. I'd rather spend a month getting failure modes right than launch fast and patch later. This isn't perfectionism. It's understanding that in infrastructure, reliability is the feature.

I avoid systems that require constant intervention to function. If it needs someone watching a dashboard 24/7, it's not automation. It's a new job.

I don't build for hypothetical scale before real usage. Premature optimization is a trap, but so is architecture that collapses the moment you exceed ten users.

I stay away from projects where the success metric is "impressive demo" rather than "still running six months later." Investors and stakeholders can be dazzled by demos. Reality cannot.

If you're working on infrastructure, energy systems, or anything where reliability matters more than features, I'd like to hear from you. Get in touch.