Introduction to the TUM-CSA Mentorship Program

Neil Wu is a PhD Candidate at the Department of Aerospace Engineering at the University of Michigan, where his research focus is on developing multifidelity approaches to multidisciplinary design optimization, particularly applied to aerospace engineering.

Multidisciplinary design optimization is a strategy that applies mathematical optimization techniques to practical engineering problems. By connecting simulation models to optimizers and computing gradients efficiently, complex systems such as full aircraft configurations can be optimized numerically, considering both the aerodynamic and the structural design.

Multifidelity design optimization is a strategy that can reduce the high computational cost in cases where the high-fidelity model is too expensive to use directly in optimization. This optimization strategy aims to handle the high-dimensional problems commonly encountered in industrial settings. Furthermore, it can also accommodate to arbitrary analysis fidelities, directly handle multidisciplinary problems, and provably converge to the high-fidelity optimum. By leveraging cheaper but less-accurate models, this optimization strategy can obtain the same optimal design but with significantly less computational cost.

Neil holds a bachelor’s degree in Engineering Science from the University of Toronto and a master’s degree in Aerospace Engineering from the University of Michigan.

Since its inception in 2009, NASA's Kepler mission has been collecting rich datasets for stellar observation. This dataset, combined with new data from the more recent TESS mission, are commonly used for identifying exoplanets via transit detection. With these datasets, astronomers have created multiple neural networks to automate the tedious process of labelling exoplanets from raw transit data. Notably, we look at AstroNet-K2, a convolutional neural network created by Andrew Vanderburg et al. which identifies exoplanets from Kepler data using previously labelled exoplanet data as training data.

Rong Yu Ming (aka. Ricky) is currently a software engineer at Google. He has spent much of the last 6 years building data systems architecture as a career. He is currently working with spacelab.space in replicating and building upon the works of Dr. Vanderburg. In this talk, Ricky will explore the process of preparing data as input for exoplanet detection via neural networks, as well as diving into some of the specific features being fed into the neural networks used for this purpose.

Josh Reid is a Senior Data Scientist at the ecommerce branch of Canada’s largest grocery distribution and retail company, Loblaws Ltd., where he is responsible for ensuring the search relevancy of online products by creating models to generate relevancy scores for queried searches. He also creates demand and inventory forecasting SARIMA models to determine optimal online assortment in order to guarantee non-stockouts. Furthermore, Josh is a veteran start-up founder of DropLab Inc a company that built custom automation equipment for laboratory research, including robotic and microfluidic systems. Josh holds a Bachelor degree in Nanoengeineering and a Master Degree in Computer Science both from the University of Waterloo in Canada.

The innaugural TUM-CSA Kick off meeting where we discuss the founding of the TUM-CSA, our mission statement, and planned events for the upcoming future.