Training modern AI systems is becoming increasingly expensive. The largest models today require enormous amounts of electricity, and if the trends continue, the models of tomorrow will require even more. There are two possible ways to reduce the energy costs, either by optimizing the hardware, or by optimizing the software. 

Amirhossein Zabihi’s doctoral journey shows how successful collaboration between research and industry can go hand in hand with a balanced schedule that enhances wellbeing. Zabihi has done his doctoral research in collaboration with Tampere University and Wärtsilä. Wärtsilä is a global leader in innovative technologies and services shaping the decarbonisation of marine and energy.