My Internship Experience – Fuzzy Logix

Does big data excite you? Do you dream about going through mountains of data to find out useful information? If yes,  then read on as Tathagat Gupta recounts his awesome experience in Bangalore working with Fuzzy Logix



Big-Data What?

Before we begin let me give you a sense of what Fuzzy Logix does. Suppose yours is a company which generates or obtains a large volume of data. You wouldn’t want to just sit on this goldmine of data. What if even with the hardware capabilities of today it takes say 2 days to (1) obtain the data from the data warehouse and (2) run the desired analysis. For many large corporations, inferences obtained from a 2-day old data are already obsolete and cannot be used to improve company performance. So instead of this 2-step process, Fuzzy Logix advocates taking the analytics to the data warehouse, hence, saving all the data transfer time which forms a majority of the total time. Fuzzy Logix provides analytics solutions to such corporations by having all its big data analytics tools sit in the warehouse itself. Its flagship product DB Lytix offers hundreds of functions (e.g. various data mining functions, statistical functions, financial analysis functions and more) on various platforms such as Teradata, Aster, Oracle etc. Use of in-database analytics for big data has significantly brought down the analysis time from days and months to a few minutes and sometimes even seconds.

Selection Process

The selection process wasn’t tedious. There was the usual resume shortlisting followed by 1 or 2 rounds of interviews. As far is preparation is concerned, there isn’t any specific skill or area you are expected to master. You are expected to be thorough with your projects and courses especially those related to statistics or computer sciences. Revision of basic courses such as ESC101 and Prob-Stats are an absolute must. Other than that you are expected to be good at puzzle solving, so practice that from various online resources. If you’re inclined towards data science and statistics and you have some points in your resume to back that interest then you have a great chance of making it. That said, it is important to mention that they do not necessarily demand a background in programming or statistics or so was the case last year.

image01

The Experience

My internship experience was satisfying and fun-filled, to say the least. Satisfying in terms of the tasks we were offered and our ability to learn and complete those with a decent level of competence. Fun-filled because of the very open company environment, an amazing set of co-interns and the various perks associated with being an intern (treats !!). There were 12 interns in all, 5 from IIT Kanpur and 7 from IIT Roorkee.

In the initial few days we got acquainted with all the essential softwares and tools we would be requiring throughout the course of the internship. My work was based on DB Lytix for the Teradata Aster platform. Being a completely new platform for me, I had to go through the various commands and syntaxes for the SQL queries on this platform. Once we all gained some proficiency in writing code on our respective platforms we were handed tasks in teams. This was done in a very thoughtful manner keeping our respective backgrounds in mind.

Since I possessed some background in Machine Learning and Statistics my team was assigned to work on various clustering algorithms- K-Means, Fuzzy Clustering and Hierarchical Clustering and propose improvements in the clustering scheme. For those familiar with K-Means clustering, let me explain what I did in some detail. The K-means clustering makes random initialization of cluster centroids. So on different runs of K-Means on the same dataset we can expect variations in the final cluster results because of different initializations. My primary task was to develop heuristics that would suggest the best clustering results among all. I also got a taste of function development as we implemented an advanced K-Means++ routine. That required a bit of C++ coding but for someone who isn’t exceptional at coding, it was well within my abilities. The best part was to run my code on datasets, see it all in action and create real value from raw data. More than the code, it was the process and the research that went behind it that left me fulfilled.

Work Environment

The team at Fuzzy Logix is small and close knit. In recent years they have made great progress in terms of customer acquisition and are aggressively expanding. The environment in the office is quite comfortable and easy-going with everyone working with great enthusiasm. The employees were all very approachable and were ready to help whenever we needed it.

All the interns were offered stay by the company in a guest house near the office for the entire duration of the internship. When a group of people stay together for a time as long as 2 months, there are a lot of great memories to take back. Apart from lavish treats by the CEO and the employees, we utilized the weekends well with all the activities Bangalore has to offer (Yes! Indiranagar, Koramangala and Church Street!). We even took a weekend road-trip to Ooty and Mysore!

In conclusion, I would recommend Fuzzy Logix to people who want to get introduced to big data analytics in a setting conducive to learning and not plain labour. The fun aspect automatically comes along when you have such an environment.

Written by Tathagat Gupta