Software release estimator

How do I start with eBay- my passion and motivation. I did my both summer internships here: One at PayPal and other is at eBay itself.  This is the first place where I saw myself growing and leveraging on my tech skills. In this post,  I will walk through the exciting project I did as a part of my Masters.

                 In today’s global industry, software development has become complex and demanding. Users want content management systems to be more friendly and comprehensive enough to exhibit all features. In order to survive in this stiff competition, e commerce companies are trying to revise their business models and making it robust at a same time.  We roomies (Mihir & Devarshi) have worked on a project called Release Risk estimator for eBay- an auction giant in Silicon valley.

The operations performed by this system are described as follows:

• When the application is deployed, the system will ask users will to select the current release from drop down menu.
• When the user selects particular release, depending upon the current system date and release start date, the system will determine the phase. Typically we can put n number of phases like release schedule planning, feature look up,  code analysis, code merging from QA to Prod environment etc…
• After determining the phase, the system will display corresponding phase questions.
• The user will answer general and current phase questions based on their best knowledge and practical judgment.
• The users can also look for additional release help from the help menu displayed on upper right corner.
• After selecting all answers, the user will click on Find Risk to calculate the risk score.
• Based on the user’s selection for all questions, the system will determine the probability and calculate the final risk score.
• After counting the risk score, the system will determine its intensity (Low, Medium & High) depending upon the range of risk

In order to use this tool effectively, every question should have some measurable value associated with it and emerged based upon their answers.  We have used Decision tree algorithm to explore each possible node by traversing from root to leaf.  This project gave us insight that any tiny piece of software module goes  n number of iterations before going to production. Of course, this is one of the ways to look at it but each has to customize its quality improvement activities based upon business model and user interaction with the system. 

We found ourselves very lucky to grab an opportunity to work in azure of eBay…..We will specially like to thank our prof  and mentors for the same.

Here is a  simple snap shot of quantitative approach practiced by the system:

 

Folks, I highly recommend you to work as a intern at eBay. You will get enormous opportunities to see your self growing personally as well as professionally… Cheers to all memories I took with me from eBay…