Having enjoyed learning about both computer hardware and software, Akshat graduated from NUS Computer Engineering in 2016. Right after graduation, he worked as a research assistant at NUS, hoping to accelerate neural network inference on power constrained devices using low cost FPGA SoCs. He developed a software framework to automatically optimize convolutional neural networks to balance loads across the heterogeneous processing elements on an FPGA SoC.
Ten months’ worth of research later, Akshat joined Transcelestial, taking the reins of the vision system needed for the Pointing, Acquisition and Tracking System, Akshat has since been developing the pointing software for the Centauri device, and helping out with the embedded systems.
Do note that the recording for this session is split up into 3 parts, of which only the first is linked here. The rest can be found at our youtube channel, located here.