Interests
As an undergraduate student, I've been exploring several fields in order to understand where my interests lie.
Here are some of the projects I've undertaken over the course of my stay at IIT Bombay. Some of these are part of
coursework while others are fun, pet projects. Feel free to explore!
Predictive Maintenance for Aircraft Engines
Predictive maintenance is the next big thing in the IoT world, and is basically maintenance, but "not too early, not too late". This was an entrepreneurial project done as a part of the course ME781 (Engineering Data Mining), wherein we were required to prepare a business model and set up the technical infrastructure for a start-up that would benefit the engineering world. We chose the aviation industry, analysing aircraft sensor data using a multitude of machine learning algorithms. These results were presented in order to effectively predict failure of engines well in advance, thereby helping aircraft operators to schedule timely maintenance. This is arguably one of my most well documented things in life, let alone projects :-p. More about that by clicking on the link below.
Camouflaged Object Detection
In the field of computer vision, camouflaged object detection is one of the most challenging tasks. We took a stab at this by implementing a search-identification module based neural architecture, achieving a commendable accuracy in the process. This was done as the course project for GNR638 (ML for RS - II), during Autumn 2020. For more details about the architecture, as well as the general implementation, have a look at the presentation by clicking on the link below. Feel free to mail me if you'd like to have a look at the code.
Football Result Prediction
This is a neural network powered result prediction model for the English Premier League 2019-20. Having analysed more than twelve seasons of head-to-head data along with an exhaustive player rating database, the model predicts football match results. The accuracy achieved of 52%, at the time of writing this, is very close to the predictions of reputed bookmakers.
This was a part of the course project for CS 419, although I extended it far beyond the duration of the course. I've also authored a detailed report on the same, if you're interested, you can view it by clicking on the link below!
Workout Classification
This project analyses a large amount of data collected from personal fitness trackers and tries to predict the activity that the user was performing. An ensemble of machine learning techniques has been used, such as Decision Trees, SVMs, Random Forests and Gradient Boosting. Testing accuracy was decent, about 80%. For more details about the data used and the algorithms employed, please do check out the project report at the link below.
Modelling a Gear Casting
This was a part of the course project for ME 206. The project involved performing simulations using the E-Foundry software to determine the optimal riser and mold parameters for the casting of several industry-grade alloy steel products. The product I chose was a gear. The project was evaluated on the basis of the effects of hotspots during casting as well as feeder aids and other remedial actions suggested. Further details and the project report can be found at the link below.
Writings
Here is some stuff that I've authored, on a pretty random range of topics.
- On the idea of the asset market as a game, and the modelling and identification of the equilibria of such a game (co-authored). Click here
- On the impression of the social location of London, from a reading of Sherlock Holmes. Click here
Other Projects
Had I been living in a world without pandemics (sigh), I would've been working in my free time at the THTF Lab at IIT Bombay, designing a fluid flow calibration system for venturimeters and orificemeters.
I was also nominated by IITB to attend a summer programme at the Technical University of Denmark during the summer of 2020, however, owing to the fact that I do not reside in aforementioned world, this couldn't happen.
I usually try out many small, fun projects - particularly during this extended break - so do check this section for updates. You can also look me up on GitHub (using the link at the bottom of the page) to check out some of the other interesting things I do!