I am a third year Ph.D. student in the Department of Industrial and Enterprise Systems Engineering at the
University of Illinois, Urbana Champaign advised by Prof. Carolyn Beck. Prior to UIUC, I was a Lecturer
in Department of Mathematics at COMSATS Institute of Information Technology. Before that I completed
Master's degree at UIUC where I worked with Prof. Xin Chen and Prof. Michael Lim.
My research interests lie in the field of scalable machine learning and distributed systems. My current project focuses on distributed spectral
- PhD Student, Industrial Engineering, University of Illinois at Urbana Champaign, 2014-Present.
- MS Industrial Engineering, University of Illinois at Urbana Champaign, 2009-2011
- MS Mathematics, COMSATS Institute of Information Technology, Islamabad Pakistan, 2007-2009
- Others: LATEX, jQuery, bootstrap, CPLEX
- Shahzad Bhatti, Michael K. Lim, Ho-Yin Mak (2015), Alternative Fuel Station Location Model with Demand Learning,
Annals of Operations Research: (230) 1, pp 105-127. [link]
- Bhatti, Shahzad, Carolyn Beck, and Angelia Nedić, Large scale data clustering and graph partitioning via simulated mixing.
In 2016 IEEE 55th Conference on Decision and Control (CDC), pp. 147-152. [link]
- Shahzad Bhatti, Carolyn Beck, Angelia Nedic (2017), Data Clustering and Graph Partitioning via Simulated Mixing, TNSE (Under Review)
- Software Engineering Intern, Riverbed Technology, Summer 2016.
- Developed a root cause analysis tool for MAPI over HTTP protocol optimization for SteelHead.
- This tool collects valuable information when an exception is throw and dumps it to a file.
- This information can then help in root causing the bug that caused exception.
- The tool can be dynamically turned on/off when SteelHead is optimizing traffic.
- Research Assistant, University of Illinois at Urbana Champaign, 2014-Present.
- Developed and analyzed scalable spectral clustering algorithm based on resource diffusion on a graph.
- Showed both analytically and empirically that the algorithm accelerates spectral clustering without sacrificing accuracy.
- Currently working on distributed spectral clustering.
- Teaching Assistant, University of Illinois at Urbana-Champaign, 2014-Present.
- Conducted weekly discussion sessions and occasionally delivered guest lectures.
- Lecturer, COMSATS Institute of Information Technology, Islamabad Pakistan, 2012-2013.
- Developed course material for Linear Algebra and Introductory Probability and Statistics.
- Taught nearly 500 students over the course of two years.
- Designed and implemented simpler versions of Hadoop Distributed File System and a MapReduce framework for batch processing. Each file is replicated at three nodes to guarantee fault-tolerance for up to two simultaneous node failures. When a node fails, other node replicating the files stored on the failed node are responsible of replicating the files. Failure detector is implemented by simplifying SWIM protocol. The user can login in to anyone of the machines and store files in the system. She can also run multiple MapReduce jobs from anyone of the machines. This machine becomes the master and allocates map and reduce jobs to other nodes in the order the jobs are received. The user can provide the number of nodes she wants to use for the map job and the reduce job.
- Designed and implemented a TCP-like reliable protocol on top of the unreliable UDP. The reliability is provided by implementing a sliding window protocol.
- Developed a random maze generator and solver using graphs and disjoint sets.
- Developed a mosaic creator using k-d trees and hash tables. The user can provide image files to be used as tiles. He can also specify the density of the tiles, for example how many pixels are covered by one tile in the mosaic.
- Collected data from imdb.com and preprocessed it for multi-class sentiment analysis. Used SVM and Naive Bayes from scikit-learn library to classify the sentiments of reviews into four classes to capture sentiment polarity as well as intensity.
- Computer Science: Data Structures and Programming Principles, Software Engineering, Distributed Systems, Communication Networks, Operations Systems Design, Machine Learning, Spectral Graph Theory, Network Algorithms, Cluster Analysis
- Industrial Engineering: Stochastic Process and Applications, Linear Programming, Simulation, Convex Optimization, Dynamic Programming, Network Analysis of Systems
- Mathematics: Linear Algebra, Real Analysis, Differential Equations, Probability and Statistics
- CIIT-UIUC fellowship for academic years 2009-2011.
- Tuition waiver and a living allowance for the duration of MS Mathematics.
- Tuition waiver throughout BS Mathematics.