Through my time as a freshman at Purdue, I've collaborated on several computer vision teams that utilize both morphological and machine learning-based solutions. I have a strong understanding of how data pipelines work through not only the operational but also the training processes when constructing a custom model. Through my experience with Purdue clubs, I've seen the good and bad in team management and overall club bureaucracy, and have learned how to communicate with team leaders and my teammates effectively and clearly.
Building computer vision based drone targeting system using combination of online and custom training data that was algorithmically labeled. Optimized model to run on embedded systems using knowledge distillation techniques.
Further development and completion of CV based lane tracking algorithm using a combination of machine learning and pure CV/Morphological approaches to determine location of lane lines given unordered data.