I worked for 2 years at Marvell Technology Group, with the former Cavium team. We worked on many-core processors, and I helped design and improve our distributed cache system for 64+ core processors.
At Marvell, I learned a lot about queues and retry mechanisms, tackled tough timing optimization issues, and worked on gate-level debugging and ECO’s. In addition, I took up complete rewrites of some of our critical scripts from Perl to Python, which would be more maintainable for the future.