Presenter

Sanjif has a Masters in Mathematics and Physics from Warwick University with a focus on Theoretical Statistical Learning. In addition to a foundation in theoretical ML, he has practical experience training/building/deploying models from resource constrained embedded systems to production at scale in the datacenter with custom hardware accelerators (Groq LPUs, GPUs, TPUs, FPGAs). At Groq, Sanjif has worked on porting converged AI-HPC applications to the Groq LPU and bencmarking performance to guarantee competitive performance. His focus is primarily on scientific applications including GNNs for molecular chemistry (DFT) simulations, feed formulation for livestock, exoplanet detection, recommendation systems and so on.