About Shubham

I am a 20 year old fellow from India and came to Singapore to pursue my Bachelor's degree from NUS 4 years ago. I have actually graduated and am now working as a research engineer at SeSaMe Research Center in IDMI (http://sesame.comp.nus.edu.sg/). Since I am no longer in school, I get a lot of free time :D I wanted to work on this project because I think there would be interesting research problems to solve? I want to pursue a PhD in a year or two and so if I can work on an exciting problem, achieve good results and publish a paper in a good conference, it will really boost my PhD application to a good school.

Actually, I was interning in the SMART center (Future Urban Mobility project) in the same building in the summer of 2013 and then continued to work part-time for around a year (took it up as an Independent Study Module for one semester and worked part-time for another semester). We were doing prediction of taxi arrivals and congestion by using open event (concerts, etc.) data and weather data mined from the web and correlating it with some GPS logs given to us by Comfort. We did write up a paper once the project was over and submitted it to PerCom but they rejected it by 1:2 vote during the reviewing process. We have modified and submitted the paper to a journal now and are awaiting the outcome.

My final year project was on automated essay scoring. I was mainly trying out NLP and IR techniques to mine the essays for features and then doing machine learning using support vector machines. We achieved a high degree of correlation between our automatically assigned scores and human assigned scores. Our results did not beat commercial software but it was still an achievement since all the work was done by just one undergrad using open source tools and with no prior knowledge of NLP or IR. Thus, it is conceivable that others will be able to build on the work and do better. My report can be found here (https://www.academia.edu/7340738/Automated_Essay_Scoring)

I did an undergraduate research project on multi-robot exploration and mapping for environmental sensing applications. This dealt with path finding for autonomous under-water gliders in the presence of ocean currents. I used Gaussian Processes and modeled the whole problem an an objective function to reduce entropy. The challenge was error-prone forecasts about ocean current speed and designing algorithms that will be able to adapt to uncertain environments. My report can be found here (https://www.academia.edu/3509757/Multi_Robot_Exploration_and_Mapping_for_Environmental_Sensing_Applications).

I have interned at Imperial College London in the area of re-configurable hardware (FPGAs, GPUs, etc..). I was parallelising some image processing applications like image skeletonization and systolic insertion sort on FPGAs while I was there.

Currently, I am working in the area of data mining. I maintain and develop Trendspedia (http://trendspedia.com/). It might be difficult to understand what it does from the website itself but I can demo to you once we meet if you are interested :-) On a related note, I am working on a paper on summarizing tweets based on DBPedia's ontology hierarchy.

Apart from the above research projects, I love programming and write code whenever I have time. I was a member of NUS' Inter Collegiate Programming Contest (ICPC) team. I have been to regional competitions but sadly, never managed to reach the world finals in this competition :'( I also participate in student supercomputing competitions and did receive a prize of excellence at Asia Student Supercomputer Challenge 2014 in China.

I have been a teaching assistant of CS2010 (Data Structures and Algorithms II), CS2103 (Software Engineering) and CS3210 (Parallel Computing). I am currently co-lecturing a Master's Course in SMU. It's called Cloud and Big Data Analytics. The module code is IS605. It's an informal agreement between me and the official course coordinator in that he lets me sit in lectures in his part of the course if I give some lectures and prepare assignments on data analysis using scipy + pandas + wakari.

I founded a startup when I was in school and we did received funding + incubation support for it but it turned out to be failure. It was in the area of mobile publishing. I was the founder and lead iOS developer for that but since it is no longer in operation now, I don't think it is good to waste your time giving you a link to it. I have been the CTO of other start-ups like SportMeNow (http://www.sportmenow.com/) and BigSpoon (https://itunes.apple.com/sg/app/bigspoon/id766169367). I used to contribute to some open source projects like OMPL (https://bitbucket.org/ompl/ompl) but I get much less time now.

So, to sum up, I would really like to do research part-time. I can usually find a lot of spare time which I waste on watching anime, Bollywod movies and cricket ;)