Built a prototype for a “basket” program which enables customers to bundle their essential items to a basket each month for automatic delivery. I built the database, APIs, auto-checkout pipeline, and a frontend for internal debugging and visualization.

Worked in the research group to build a pipeline to

automatically ground truth the data values on various mortgage documents based on fusing together multiple sources

of information, such as multiple Optical Character Recognition (OCR) engines.

• Built a benchmarking tool for our object detection and tracking system to benchmark the improvement in each iteration.

• Fine-tuned our existing object detection network to detect objects at night using a thermal camera. Responsible for data collection, training, monitoring, and testing.

• Built a pipeline in opencv to compare various image feature extractors and matchers. I analyzed the effectiveness of multiple state-of-the-art neural nets for our use case compared with more classical image feature extractors and matchers, and helped integrate them into our computer vision stack.