About

Aditya Jindal is a highly motivated and skilled Machine Learning Engineer with a passion for computer vision. He is currently pursuing a Bachelor's degree in Economics at IIT-Kanpur, with a Master's degree in the same field to be completed in 2022. Aditya is a Regionalist at ICPC'20 and works as a Data Scientist at Ping Identity. Aditya has a diverse set of skills and interests, which include business, technology, trading, markets, research, and life. He enjoys talking about these topics and is always looking for something interesting to discuss. Aditya is a keen reader and loves to stay up-to-date with the latest advancements in the industry. Aditya has experience working as a Quantitative Research Intern at alpha unit technologies, where he built robust algorithmic momentum strategies for trading equities. He built ML and statistical models to predict market signals and devised momentum indicators like Relative Strength Index, Stochastic Oscillator, and Momentum Price Strength for signal predictions. He also implemented an ensemble model of Random Forest & K Nearest Neighbor with Dynamic Time Warping as a distance metric. Aditya has also worked as a Computer Vision Intern at Kamerai, where he implemented an efficient framework to detect arbitrary-shaped text in real-time on tote packages for product-id recognition. He proposed lightweight backbones, Mobilenet, and Shufflenet on top of FPEM segmentation head to get the faster inference. He improved the inference speed by 1.5 FPS on their custom Tote-ID dataset and by 6 FPS on the ICDAR-2015 Dataset. Aditya also worked on face recognition using Dlib Toolkit, where he implemented a c++ based Dlib metric learning model for producing embeddings on employee-attendance image dataset. Aditya has also worked as a Research Project member at Indian Institute of Technology, Kanpur, where he tackled the Semeval problem of Table-based Statement Verification and relevant cells selection in the table for each statement. He preprocessed a corpus of 3000 tables and used entity linking to link statements with tables. Aditya also proposed a new model CellBERT, for the task of Evidence finding from tables and achieved an F1 score of 0.65 on Evidence finding by ensembling CellBERT with their heuristic-based approach. Aditya's technical skills include Research Scientist, Data Scientist,