I'm a currently an ML engineer at Hike India. I completed my undergraduate degree with majors in Computer Science and Engineering from Birla Institute of Technology Mesra, Patna. I have worked in related fields like machine learning, artificial intelligence, software engineering, internet of things and data mining. I strive to contribute in machine learning fields as the concept and large applicabilty of machine learning fascinates my curious mind and provides a space to come up with my own ideas.
"Growth occurs when one goes beyond one's limit. Realizing that is also a part of the training."
I enjoy learning new and keep moving forward so that I could acquire as much as I could get. I consider work as an ongoing process, and I'm always looking for opportunities to work with those who are willing to share their knowledge. At the end of the day, my primary goal is to work hard and gather knowledge.
When I'm not in front of a computer screen, I'm probably reading books, thinking about robotics, playing football, reading mangas, or cooking.
Automated Medical Assistance is my research, where we developed a conversational dialogue system that is able to provide better consultation during the need and is able to answer any queries related to one’s health.
In this paper, we presented an analyses of the resource efficient predictive models, present in the machine learning field for resource constraint devices. These models try to minimize resource requirements like RAM and storage without hurting the accuracy much. We utilized these models on multiple benchmark natural language processing tasks, which were sentimental analysis, spam message detection, emotion analysis and fake news classification.
A novel character-level pretrained language model framework which utilizes the transformers and character-level language models to extract sentiment phrases.
MemSem is a multi-modal framework for sentimental analysis of memes. With the help of multi modal architecture it extract features from the input to provide a meaningful analysis for a meme, whether its offensive, humorous or non-sense. The project is published in IEEE BigMM 2020 Publications and is still in progress.
QuesBELM is a natural question answering system, based on natural language processing and ensemble methods. It works on an ensemble model of BERT-base, Albert-xxl and BERT-large. Trained on SQuad2.0 dataset, preprocessed and sampled to provide better results. It produces outstanding results and is helpful for researches related to natural question answering.
ASL-Classifier is a python wrapper object. It's a real-time based classifier, helps in detecting and classifying American Sign language. With the help of Haarscascade and opencv it detects palm in the view and uses pretrained keras model in classification. It uses opencv library to create window applications to manage settings and classify output.
Tweeple DNA is a FastAPI based project that classifies twitter-people's account into bot or human, as well as assigns probabilty for being a bot ranging from 0-1.
hiLyted is a project which clips highlights of video. It is based on the method of short time energy in audios extracted from a video. It captures the time period containing high pitch sound considering it as audience applause during the tournament. It downloads the video and audio from youtube using youtube-dl and extracts audio feature using Librosa, with the help of MoviePy it clips the time period of the highlights and saves it in a local directory.