Most companies simply collect all the data they can without much consideration for how exactly this could be used. This has become more so because deep learning makes it easier to use data that was traditionally of limited use. Most discussions on machine learning and deep learning ignores the data preparation part of the problem. However, this can take up to 60% of the time for most analytics projects. In this talk, I talk about how human computation and data centred design can enable much faster and more efficient learning and more powerful AI systems.
Vaisagh is a founder and CTO at impress.ai. impress.ai is a productivity tool for enterprise recruitment teams to allow them to identify great candidates faster, more accurately and without bias using an AI-powered Virtual HR Assistant. Before impress.ai, he completed his PhD on developing agent based models for understanding human crowds from Nanyang Technological University and subsequently worked as a Research Fellow at the NRF-funded institute- TUM CREATE, where he lead a team working on building and using city-scale traffic simulations to study different aspects of urban mobility.