About me

Hello ! Welcome to my page. My name is Pratham Desai, I hail from India. I am a first year graduate student at University of Wisconsin-Madison in department of computer science. You can look at my academics, research projects and internships in the following sections. Apart from computer science, I love watching movies, listening to songs and playing keyboard.


  • Graduate Student, Computer Science at University of Wisconsin-Madison.
  • Major with Honours in Computer Science and Engineering at IITB.
  • Minor in Applied Statistics And Informatics.

Academic Performance

Examination Board/University Year Performance
Undergraduate IIT Bombay 2016 9.34/10.0
Intermediate Gujarat Education Board 2012 85.53%
Matriculation Gujarat Education Board 2010 93.85%

Scholastic Achievements

  • Awarded Special CS Scholarship from University of Wisconsin-Madison
  • Awarded Institute Academic Prize for exceptional performance among 880 students during year 2012-13
  • Awarded AP grade for extraordinary performance in the class for courses on Machine Learning, Data Analysis and Interpretation and Numerical Analysis at Indian Institute Of Technology, Bombay
  • Awarded the prestigious KVPY Scholarship (Kishore Vaigyanik Protsahan Yojana), instituted by the Department of Science and Technology, Government of India, with All India Rank 87
  • One of the 5 recipients of Aruna Lal Scholarship given by Physical Research Laboratory, Ahmedabad

Projects & Internships

Research Projects

Multi-Objective Query Optimization

Prof. S. Sudarshan
  • Worked on search strategies to find optimal query execution plan efficiently from exponentially large space, used response time based multi-objective cost model for parallel computing environments
  • Designed and evaluated heuristics for improving depth-first search by prioritizing more promising plans based on local costs and statistical estimates
Poisson Denoising of Low Intensity Images
Autumn 2015

Prof. Suyash Awate
  • Worked on dictionary learning based image denoising techniques for low SNR images
  • Formulated a minimization problem considering Poisson noise model and developed an algorithm to reconstruct noisy images using sparse non-negative combinations of trained feature vectors

Key Academic Projects

Video Stabilization
Spring 2016

Prof. Ajit Rajwade
  • Implemented an application to determine a sequence of translational+rotational transformations between SIFT feature points of consecutive frames in a jerky video
  • Used RANSAC algorithm for robust estimation of the model parameters
Smart Music Controller
Spring 2016

Prof. Kavi Arya
  • Developed a device with Raspberry Pi for auto-controlling a music player based on content and surroundings using music-speech discrimination models
  • Incorporated adaptive volume control based on the amount of conversation going on in the background, automated switching of radio channels when music is not being played
Stock Price Predictor
Spring 2016

Prof. Ganesh Ramakrishnan
  • Applied machine learning techniques to predict the direction of stock price movement across the day
  • Used SVM and neural networks for classification, Experimented with features like recent history, bond rate, momentum of the market to find their effectiveness
Text-to-Speech Conversion Using Neural Networks
Spring 2015

Prof. Pushpak Bhattacharya
  • Worked on feedforward neural networks with error back-propagation through multiple hidden layers for converting graphemes to phonemes
  • Implemented adaptive learning algorithm and compared various text-encoding schemes for improving convergence rate while learning the weights on sigmoid neurons
  • Achieved 89% accuracy on CMU dictionary dataset with 5-fold cross validation
Compiler for a Subset of C Language
Spring 2015

Prof. Amitabha Sanyal
  • Implemented a compiler for a C like language to output assembly code using Flex and Bison
  • Supported arithmetic operations, function calls, recursion, function overloading and decision constructs with lazy evaluation
  • Implemented Sethi-Ullman algorithm for efficient register allocation during evaluation of arithmetic expressions
Analysis of Replacement Strategies for Page Level Buffering
Autumn 2014

Prof. N. L. Sarda
  • Implemented LRU and MRU page eviction strategies for the buffer in a toy database and analyzed their impacts on buffer hit rate
  • Compared these strategies based on their performance on three different join algorithms
  • Evaluated the improvements in query execution time achieved by rewriting plans using basic equivalence rules on the query expression tree
Hotel Booking Application
Autumn 2014

Prof. N. L. Sarda
  • Developed a 3-tier web-application for hotel database management and online book- ings, ensured data consistency for concurrent operations
  • Incorporated dynamic pricing where prices keep changing according to season, trend, duration of booking and user history


Improving landmark scenario in Bing maps applications
Summer 2015

Bing, Microsoft India Development Center
  • Designed and implemented heuristics to find out duplicate entities in the neighbourhood of a point using approximate string matching and phonetic matching algorithms
  • Proposed a machine learning based solution for the problem of clustering business entities with constraints on geographic proximity and uniform distribution among clusters
  • Received a pre placement offer after successful completion of the project
Analysis and Visualization Of Genetic Data
Summer 2014

Strand Life Sciences Bangalore
  • Worked on analysis and visualization of genetic datasets
  • Developed a web based querying system for the same as an in-house application
  • Designed a database schema and indexing policies to operate efficiently on around 63.5 million variant tuples stored in dbSNP (Single Nucleotide Polymorphism database), further optimized the model to respond dynamically to frequent and complex queries

Extra curricular Interests

A big fan Of Imagine Dragons. This is the most amazing band in my opinion. For their songs, it is said that you don't like it that much at first. You listen it a couple of times and it suddenly becomes your favourite. Although almost all songs ar unique and lovely, if I have to name a few, I can say Demons, Shots, Roots, It's Time and I bet my life are my favourites.

Madly in love with A. R. Rahman's compositions, too ! I feel that there is right mixture of classical and modern music in his creations. Album of Rockstar(2008) composed by him was one of its kind with beautiful songs in soulful voice of Mohit Chauhan. Here are some links to his magnificent work: Kaise Mujhe Tum Mil Gayi, Tu hi re, Jashna-e-bahara, Jo bhi mein, Aur Ho, Phir Se Ud Chala.

There are some movies I would not mind watching again and again. The Dark Knight Trilogy by Christopher Nolan is my all time favourite. Avatar, The Prestige, Avengers, Inception, Pursuit Of Happyness, Shawshank Redemption are also masterpieces.



Email me on:


Created Using Bootstrap Scrolling Nav