Umang Sharan
- Famnom
Famnom is an easy to use, micro-nutrient and macro-nutrient tracker. The website is hosted on Heroku, and built using Django, and PostgreSQL. The iOS and Android apps are built on Flutter. The app leverages DRF and Django's rich plugin library for rapid protoyping.
Famnom offers tools to set nutrition goals, log meals, set custom nutrition preferences, and generate daily meal plans based on nutrition preferences and food availability. Family accounts are supported to share foods and recipes in the kitchen among family members.
Web URL: https://www.famnom.com
iOS app: https://apps.apple.com/us/app/famnom/id1583273562
Android app: https://play.google.com/store/apps/details?id=com.famnom.famnom
Server code: https://github.com/umangsh/famnom
App code: https://github.com/umangsh/famnom_flutter
- Dhunio
Dhunio (Offline, an old archive here) is a YouTube catalog for bollywood music. The website is hosted on AWS, and built using Django, MySQL, and Solr. The project involved several challenging pieces. Since there is no canonical feed for bollywood music, we ingested and reconciled data from multiple feeds to generate an authoritative catalog. The catalog is browsable by year, cast, etc. The data crawlers and metadata reconciliation pipelines were written in python. This was a fun exercise in building and deploying a large scale service using Django, Solr, AWS, S3, python, etc.
- vSketch
I released my first iPhone application called vSketch in 2009. You can get the app here (No longer available on App Store) for a very reasonable price of $0. The app is easy to use - you choose a photo from your photo gallery or click a picture from the camera and convert it either to a pencil sketch or an oil painting. The app uses canny edge detector for generating the pencil sketch, and image segmentation and bilateral median filtering to convert the photo to a painting. The app has been downloaded more than >50,000 times.
I extended the algorithms used by vSketch for Non Photorealistic Rendering (NPR), to videos. The resulting effects are very similar to the effects seen in the movie A Scanner Darkly. You can find all the NPR videos on my YouTube channel.