About me

I am Pranav Jangir, a Computer Science MS student at NYU Courant. I am intersted in algorithmic game theory, machine learning, privacy and cryptography.

What am I busy with nowadays?

I thought you’d never ask. I am currently focused on designing social welfare maximising algorithms for seminal problems in Game Theory; specifically in the burgeoning field of learning augmented algorithms, that attempt to move away from the pessimistic worst case guarentees for algorithms.

I am also fascinated by the use of machine learning in healthcare. Currently, I am working on a research paper that focuses on a specific technique called fast subset scan, which is used to identify affected regions in spatio-temporal datasets when prior information is available. The main objective is to efficiently scan through all the subsets of the data based on the given prior information. This prior information can take various forms. For example, let’s consider a scenario where we already know that the regions affected by a disease are spatially connected and loosely resemble a circle.
In practice, this prior information is often not well-organized, and a straightforward search method would be prohibitively slow due to the exponential number of subsets. Additionally, storing the prior information in memory is not feasible.

I am also continuing my research on developing privacy preserving algorithms and we submitted a paper recently which should be out soon (hopefully)!

History

Software Engineer @ Google Ads Quality

Prior to joining NYU Courant, I worked at Google in the Ads Quality team where my team focused on improving the quality of ads for low resource languages like Hindi, Polish, Swedish etc. that constituted of non-trivial percentage of query traffic but having limited training data.

Cryptography @ IISc

I spent some amazing time as a Research intern at IISc, Bangalore under Prof. Arpita Patra in their Cryptography and Information Security Lab.
We worked on developing fast and secure privacy preserving algorithms for various problems of interest such as the Heavy Hitters problem. We also explored some oblivious RAM algorithms and some areas in theoretical cryptography like order preserving encryption.

Intern @ Tower Research Capital

I interned in the Post Trade Division. Our team’s project was to validate financial data based on past transactions.

Brokers provided financial data in pdf format. Extracing information from the pdf files required writing a separate parser for every format that was done by contractors and took about a week.

I built a semi supervised machine learning pipeline that parsed any new pdf file with minimum human supervision. The effective time to parse a file reduced from a week to about a day.

Math Undergrad @ IIT Guwahati

My course involed more applied math than pure math. My courses included Probability (Basic and Advanced), Real Analysis (Basic and Advanced), Financial Engineering, Stocastic calculus, Combinatorics, Optimization and some other usual CS courses.

I spent most of my undergrad doing the thing I love the most, solving puzzles (more specifically algorithmic puzzles, but all types of puzzles are welcome). I took part in Topcoder, Codeforces and Atcoder contests.

I was inspired by the movie The Man Who Knew Infinity and in a bout of over-confidence decided that I should do my Bachelor’s thesis at the interesection of Number Theory and Combinatorics. I finally wrote my thesis on Partition Functions.