Currently, the systematic miniaturization and increased speed of integrated circuits are pushing electronic components to the limit. With the increase of AI in systems, the goal will be to be able to constantly optimize voltages and operating speeds regardless of the conditions, and eventually, to be able to perform predictive maintenance. In other words, the system will be able to autonomously predict a failure, or even to put itself in safety to avoid any collateral damage. The role of the researcher will therefore be to train and verify the consistency of the AI results to guarantee a failure-free electronic circuit.