Developing machine learning pipelines to analyze cellular imaging and contribute to the discovery of novel therapeutics.

• Managed several concurrent computer vision and machine learning projects for aortic disease treatment planning, leading to three new software tools, a novel lab research direction, and several research abstracts and publications.

• Optimized 3D medical imaging neural network pipelines using MONAI and Pytorch, achieving a 100x increase in training epoch speed via parallel GPU training and data caching strategies, reducing processing cost and time.

• Developed custom data labeling tools, automated label validation methods, data augmentation strategies, and a hyperparameter search to create a 700 GB medical image dataset and train a U-Net model to segment arteries.

• Designed compact, low-precision neural networks and digital filters optimized for deployment on edge devices.

• Created a versatile Monte Carlo simulation codebase, enabling performance evaluation of RTL designs tailored for digital filtering and machine learning, resulting in new optimizations and higher application accuracy.

• Developed neural network training pipelines that reduce model complexity and size through quantization and pruning, resulting in designs that utilize less 10% to 35% less resources while maintaining high application accuracy.

• Published seven first-author papers in leading IEEE and ACM venues like DATE, ICCAD, TODAES, and BioCAS.

I delivered weekly discussion lectures and developed homework/exam questions, among other responsibility. For these tasks, I started with clear learning objectives and worked backwards to develop course material. By putting student learning needs first, I was able to make efficient use of class time which ultimately led to emphatic student reviews and a 4.9/5.0 instructor effectiveness rating at the end of term.

Following one semester of teaching, I was fortunate enough to advise two former undergraduate students in research. My advisees and I had complementary expertise which led to a productive collaboration. Ultimately, our efforts led to an IEEE BioCAS conference publication that applied stochastic circuits (my expertise) to hearing aids (my student's expertise).

The importance of pedagogical techniques and philosophy extend beyond the classroom and given this belief, I regularly attended workshops targeted at improving specific aspects of teaching. Through these workshops and other programs, my teaching philosophy has evolved into one that emphasizes motivation, critical thinking, and community.

Courses:

Fall 2017: EECS 203: Discrete Mathematics

Winter 2018: EECS 270: Introduction to Logic Design

Winter 2019: EECS 478: Logic Circuit Synthesis and Optimization (Graduate Level)

Winter 2020: EECS 478: Logic Circuit Synthesis and Optimization (Graduate Level)

I worked alongside other tutors to run drop-in physics tutoring sessions. I tutored all general education physics courses as well as the physics major main track courses up to modern physics (a course that serves as an introduction to quantum mechanics).