You will be responsible for compressing & quantising neural networks, develop tools and methodologies to perform platform specific quantisation with minimal loss in accuracy. Develop new ideas for machine learning within Pytorch. Optimize existing implementations within Pytorch. Research & implement distributed learning strategies. Enhance multi-gpu and multi-node capabilities in Pytorch. Interface with research team to understand neural network algorithms. Research and develop compression strategies for the target devices. Document work artifacts and ensure alignment with internal development processes.