About Me
Sharma
Tech Enthusiast. Love for Artificial Intelligence. A desire to change the society by contributing to the on-going Artificial Intelligence revolution. I like the Hustle!
My work predominantly revolves around adding intelligence to automotive domain. I have helped converting various projects from Proof-of-Concepts (POCs) to long-term projects, mentored and lead teams from technical front. Built several POCs using AI which are converted as full-scale projects like Driver Monitoring System (ADAS) on the Edge, Car Parking Solution for Multi-Storey Parking, Social-Distancing and Face Mask Detection Solution for COVID-19. Inducing Photo-realism (urban-scene) to Synthetic Data using GANs. Developed a solution for GDPR compliance using Computer Vision & AI. Using AI to improve road safety by identifying near-miss incidents at a pedestrian crossing in London (Lambeth). Published in London News.
March 2019 - PresentI am working with Prof. C.V.Jawahar at the Centre for Visual Information Technology (CVIT) Lab on Document Image Super-Resolution. Part of my work on how Document Image Super-Resolution is different from Natural Image Super-Resolution will be submitted to CVPR 2019. We are using Deep Learning to solve the document image super-resolution problem.
August 2018 - February 2019Worked with Dr. Arnav Bhavsar & Dr. Aditya Nigam at the Multimedia Analytics, Networks and Systems (MANAS) Lab on applying Deep Learning to Medical Imaging & Biometrics domain. Part of my first work on generating a detailed 7-Tesla Brain MRI from a low quality 3-Tesla MRI scan is published in ICIP-2018 as an equal first author contribution. Our 7-T MRI reconstruction algorithms outperformed several benchmarks in terms of reconstruction performance and computational efficiency. In my second work, we proposed a novel CNN based multi-decoder framework for reconstructing 7-Tesla brain MRI from 3-Tesla brain MRI along with segmentation results is published in DLMIA-2018 workshop (held with MICCAI-2018) as an equal first author contribution.
December 2017 - July 2018Contributed as a speaker at National Conference on Computer Vision, Pattern Recognition, Image Processing and Graphics (NCVPRIPG 2017) on a deep learning session and at AMU (Aligarh Muslim University) for a five days workshop on deep learning.
December 2017 - PresentAt DataCamp, I write articles related to Deep Learning with a combination of both theory and practical implementation.
November 2017 - PresentAnalyzed data for neural networks and classified catalog data using deep neural networks. Used image processing techniques to extract features. Also worked on Data Analysis using Dataframes, wrote unit tests for production ready code. One of my task was related to NLP using SVM to predict true & deceptive hotel reviews, and deployed the model using Sanic as backend.
August 2017 - November 2017I taught underprivileged children at a NGO “Cry India” for almost 7 months, where I taught Hindi, English, Physics and Mathematics to the Grade 9 and Grade 10 students.
September 2016 - December 2016Identified, evaluated and scaled-up voice lines systems to enhance customer calling experience and enabled business to take better decisions. Customer Support & Satisfaction – Achieved high level of customer satisfaction through patient, methodical approach to problem resolution. Received performance points for appreciations from manager and users.
June 2015 - August 2016
In this project my model had to detect given a review whether it is classified as a true or a deceptive review. I achieved an accuracy of around 86% with naive bayes classifier and 88%-89% accuracy with SVM on the training and test set. The predictions of my model were similar to the predictions made by the tool reviewskeptic.com/ which was developed by Cornell. Finally, I deployed my model using a back end similar to flask in python called Sanic.
© 2021 Aditya Sharma
