Loading Projects...
Flikcer is an award winning novel website and chrome extension to resolve photosensitive epileptic content in videos. It not only provides a measure of possible triggers of epilepsy but also provides a safer version of the video by eliminating all frames that could be harmful.
Savvy is a mobile application and widget that creates awareness about lesser known challenges the world faces today, by providing visually pleasing infographics, facts and statistics. It also works as a platform for NGOs to meet donors and possible volunteers.
Plothole is an app that can automatically detect, geo-tag and report potholes. The Government of Delhi implemented the app in Delhi's largest road campaign, where 50 MLAs, along with PWD engineers, took part in road surveys. A total of 1,897 reports were filed.
CASI, the acronym for Cardiac Arrest Stress Indicator, is a routinely-disposable patch designed to detect, ergo, prevent any affliction, specifically cardiovascular diseases, using elementary chemical properties and contemporary technological concepts. These integrate to provide accurate insight to monitor the stress level of an individual.
BasketAI is a sales forecasting deep neural network based on FBProphet and a rolling average network. The model has been created to be computationally efficient, and is now being used by BigBasket, India's largest online grocery, to process sales of over 22,000 products within practical time limits.
We built a platform to provide actionable insights to government agencies and social workers to help them understand the impact of fires, and see past and forecasted fire tracks to be better able to allocate resources. Phoenix is an intuitive easy-to-use open-source dashboard equipped with a number of novel machine learning and deep learning models to ingest data in real-time.
2018: A mathematical model for fairer ranking systems based on cumulative binomial distribution and linear regression
2019: Predictive climate change model using multivariate regression and solution-mapping with superpixel segmentation and K means clustering