|
Before joining the University of Maryland as a PhD student, I worked at Minekey Inc for almost 4 years(May 2006 - June 2010). Minekey (TechCrunch, CrunchBase, GigaOm) was incubated at IIT Kharagpur where I did my undergrduation. I worked as a student researcher during the early stages of Minekey. Later Minekey got funded by NEA IndoUS Ventures, and I joined the India development center full time. Following are few of the selected projects I did while at Minekey.
Back end architecture and development for Twezr (TechCrunch, CrunchBase)
Description : Twezr is a universal Inbox; it's a one-stop destination for all your email and social network accounts. Currently
it supports Gmail,Yahoo,AOL,MS Exchange, Facebook, Twitter and any other IMAP based email accounts. I co-owned prototyping the back-end functionality and architecture. I worked on crawling,analyzing,structuring information from email/social network accounts; designing algorithms to merge(same person accross platform being clubbed together, making it easy to track people) and rank the contacts for those accounts; impelementation of Tomcat based webservice API to talk to the front end. The database we used was a blend of CouchDB and MyQSL storage units distributed and replicated over multiple instances.
Minekey facebook Application and Minekey Website (CrunchBase)
Description : iThink is an online discussion network where one can share his/her opinions in a commnon forum and initiate constructive discussions. As a back-end engineer I was closely involved with the system design and scalability of the application. Memcache, Mysql replication ,Sharding, Push based user information queueing were the key techniques we deployed. We had this application on multiple social network namely Orkut, Facebook, Myspace, Friendster and Hi5.
I led the development of the full fleged iThink website as well.
Minekey Recommendation Engine(ChrunchBase)
Description : It was a recommendation service for bloggers and it served news/blog articles to the readers based on their reading habits. I was responsible for extracting all the informations from online resources , analyzing the them and then recommending articles to the users based on their explicit(favorites) and implicit(reading history) interests. Some of the key challenges were implementing crawler to fetch data; using categorization, clustering, language processing to extract concepts; modelling users based on their reading habits to personalize news.
|