Behind Nordic Gradient

I'm John Dahlberg, a computer vision engineer based in Stockholm. My previous work includes computer vision for self-driving trucks at Scania and a patented deep learning model for brain radiosurgery planning at Elekta. I hold an M.Sc. in Systems, Control and Robotics from KTH, with a thesis on explainable AI.

I started Nordic Gradient to take on computer vision problems I find worth solving, from the first research to working code. Outside client work, I build things for my own use, like home automation and a detector for Nordic wildlife.

Why the name

Nordic is where I come from and where I work: Stockholm, one of Europe's centres of tech innovation. It also stands for how I like to build things, in the Scandinavian tradition of design: simple, and made to last.

Gradient comes from gradient descent, the method neural networks learn by: small steps downhill on a landscape of errors, until the error is as low as it gets. The logo draws both halves, an aurora in the arc of the G and the descent below it.

Elsewhere

• CV (PDF)
• LinkedIn
• GitHub
• Kaggle