About

Aditya Bingi is a Senior Computer Vision Researcher at PureCode Software with a passion for understanding and engineering intelligence to solve real-world problems. He deeply cares about the meta-learning paradigm and how it can help artificial agents learn from limited data and generalize to novel scenarios. Aditya is currently working on meta-learning in computer vision and reinforcement learning areas. Aditya has 4.77 years of relevant experience and is a skilled Research Scientist, Data Scientist, and Deep Learning practitioner. He has a strong background in computer vision and has worked with technologies such as Tensorflow, Keras, CNN, OpenCV, and Pytorch. Aditya has also worked with AWS, Hyperparameter, Normalization, Dice loss, test, Neural Networks, and Dagger. In his current role at lumiq, Aditya has built an end-to-end invoice information extraction system for one of their clients. He ideated and designed a three-step pipeline for invoice information extraction, cleaned and processed noisy data of scanned invoice images using OpenCV lib tools, and worked with EAST text-detector and open-source Google Tesseract-OCR on their invoices data along with their own created synthetic data. Aditya also built their own OCR system from scratch based on Facebook's Rosetta paper in Tensorflow, tweaking the ResNet-18 architecture and modifying it to use as a backbone for the text recognition task. Additionally, he added a Bi-LSTM layer on top of their CNN-backbone, which improved their OCR character error rate to 2%. The beam-search algorithm is used to recognize incorrectly identified text, further enhancing their OCR system. Aditya has also designed heuristics and rule-based algorithms to extract information from recognized text. Aditya has previously worked with the National Institute of Hydrology, where he evaluated the impact of reservoirs on groundwater recharge and vegetation in the Bundelkhand region using Python and GIS software platform analysis. He has also worked with the Geological Survey of India as a Summer Intern, where he carried out aerial photograph and satellite image interpretation of an area pertaining to two different blocks in Prakasham, Kurnool, and Anantapur districts of Andhra Pradesh. Aditya holds an Integrated M.Tech degree in Geological and Earth Sciences/Geosciences from the Indian Institute of Technology, Roorkee. He also has a degree in Intermediate from NRI Academy and completed his high school education from Sri Nikileshwarananda High School,