Computer Sciences Dept.

Rohit Koul

Graduate Student

  Home -->  Misc 


  Research Interests
  Patents and Papers



In keeping with the recommendations of the TPILB-Project, this website features a blank page 

! FAQ..

Some !frequently asked questions about this site are here

Tech Log

The purpose of this log is to put together a list of some of the other non-academic projects I have undertaken in the past couple of years

BITSAA 501(C)(3) Non-Profit

Update     I have since moved into Strategy & Operations, but I still help in design decisions.
Update 2:: The projects as under are kinda outdated.

BITSAA is a charitable Non-Profit, registered in NJ (closely asscoiated with my undergraduate institution BITS Pilani) run by 300+ volunteers in varied capacaities for different teams. I handle the technology side of it, the alumni portal, design and manage its custom applications for various initiatives and help take technology decisions.
I joined in Feb/Mar 2010 as a volunteer with the PHP knowledge of having written <?php echo "hello world" ?> only. Since then I have had to learn couple of new tools and techniques and apply them to get things done for the team. Some recent ones are as under:
  • Twitter integration using OAuth - the idea was to make users change their twitter background image by a single click.
  • Authentication and webscraping using cURL and PHP.
  • Webcrawling and scraping data using wget
  • A PHP based project management system
  • A simple system to send credentials to various members securely using emails.
  • Bulk-mailer to work-around the limits set by the company hosting our infrastructure.
  • Credit-Card fraud prevention checks
  • Captchas
Oracle Social Network:
In 2007-08, While at Oracle, one of the nice things (I think!) I did (with 2 others) was to develop an internal social networking platform - we called 'iMingle'. The goal was to create a non-email based channel for people to collaborate, come out of respective product silos, get to know each other, find common interest groups and have fun. I had just discovered the power of Ruby on Rails (ROR) and one other motivation was to see what all we could do with it.

Three of us (Rakesh K (who is now pursuing his PhD at UIUC), Varun Khurana (who, I believe, is pursuing his MBA from Cornell) and I (who as you would have known by now, is stuck at UW-Madison :)) ) came together under the guidance of Suresh Srinivasan (CMTS, Director OHS, OJVM etc etc), to build this tool. We did most of it during our free time, over weekends and staying very very late at office during weekdays. Yes, we didn't have a life during that period, but we were driven by creating something that we knew would have an impact. Besides, we were calling the shots, deciding what feature should go in and what not, so the feeling was quite different from our day job. (which btw, I did like, thanks to my team and my incredibly talented manager Deepak Ramakrishnan).

Our 0.1 version had profile pages, friends, groups, very very versatile people search (thanks to ferret), blogs and forums. Some people liked it, some didn't (meh!) but although our target audience was 400-500 Oracle Application Sever employees, it grew to span groups across multiple centers in Oracle India and US and served more than 1100 users and 100+ groups within a few months of its Beta release.We, I believe, managed to create some buzz and got in some volunteers to help build some more features like polls, bookmarks.

I am not sure what the status of the toy is now, but It did motivate a bunch of other folks to build similar products within Oracle India. Slides from one of the many many presentations that we gave are here. It has a couple of screenshots of iMingle (yes, we sucked at colors. Deciding the right shade for Orange/Green for the top bar took Varun and I more than 3 hours of experimentation and we were still not happy at the end of that frustrating exercise. :) ). But, heck, it worked. According to me, It just needed someone with an eye for pleasing color combination. (some others loved these colors btw!)

Aside: In hindsight, now that I am studying Distributed Systems in grad school, I can see a bunch of mistakes we did. The site was definitely not designed for scale and we were mad about consistency!! (which as I am learning does not gel well with high availability and partition tolerance.!). Also, we probably didn't need a full fledged database system. some NoSQL stores may have performed better.


Copyleft and Copyright © Rohit Koul, 2009-10 
Developed on a     box using