Christophe Gyurgyik is a PhD candidate at Stanford University. His interests include languages (programming and natural), compilers, and hardware-software codesign.
Previously, he worked at Google on the XLA (Accelerated Linear Algebra) compiler and the TPU (Tensor Processing Unit). Prior to industry, his researched focused on Calyx, a compiler infrastructure for languages that target hardware accelerators. This led to further collaboration with CIRCT, an experimental compiler to improve hardware design tools. The work was completed under Cornell University’s CAPRA, spearheaded by Adrian Sampson and Rachit Nigam.
In his free time, Christophe enjoys speaking natural languages (French and Italian) and rock climbing. A recent memorable adventure was the Tuolumne Triple Crown.
Officially began my PhD candidacy at Stanford University.
Left Google after two years with the XLA compiler and TPU backend team.
Received the Cloud Tech impact Awards (CTIA) for LLM serving optimizations.
Awarded a spot bonus from the VP of Engineering (Google) for improvements to large language models through compiler optimizations.
Our paper Stepwise Debugging for Hardware Accelerators won the distinguished artifact award at ASPLOS ’23.