About

Pushpendu Ghosh is an experienced Applied Scientist at Amazon, with a focus on exploring novel ways to improve the catalog quality of Amazon using CV and ML systems. With over 3 years of relevant experience, Pushpendu has a strong background in research, ML, computer vision, and software engineering. Currently, as an Applied Scientist II at Amazon, Pushpendu is responsible for developing and implementing cutting-edge technologies to enhance the catalog quality of Amazon. In his previous role as an Applied Scientist I at Amazon, he successfully extracted dimensions from product images, translated text on product images, classified and estimated relevance of browse nodes, detected restricted content in images, and resolved image-title mismatches. He has filed for 2 patents and has had 4 internal paper acceptances. Pushpendu's experience also includes an internship at Amazon, where he worked on a project to detect defects in product detail pages. He designed and implemented a multi-modal model using Xception, FastText/VowapalWabbit, and Pyspark and Tensorflow technologies. The application dataset size was 100 million products across more than 10k categories. Pushpendu earned a Master of Science and Bachelor of Engineering in Mathematics and Computer Science from Birla Institute of Technology and Science, Pilani, in 2020. He has a strong tech stack and is well-versed in research, ML, computer vision, and software engineering.