SUMMARY

As a data scientist I have 13 years of experience in statistical modeling, probabilistic programming, visualization and analysis of large data sets using a variety of tools and languages such as Python. I have led and taken part in research projects independently or in collaboration with various groups (ranging from one to more than 90 collaborators on a single project), have published more than 60 research articles in peer reviewed journals with more than 2500 citations, have taught various courses in different countries in the past two decades and have given lectures at universities and presented in conferences in Canada, US and around the world.

EDUCATION

Inter University Centre for Astronomy and Astrophysics

Ph.D. Astrophysics.

Institute for Advanced Studies in Basic Sciences

Master of Science, Physics.

Sharif University of Technology

Bachelor of Science, Physics.

EXPERIENCE

Senior Research Associate, Canadian Institute for Theoretical Astrophysics (University of Toronto)

Toronto, ON — 2014 - Present

Using advanced machine learning and statistics to analyze large datasets.

Accomplishments

Used semi-supervised machine learning techniques to simulate maps from mock catalogs of clusters of galaxies (PDF paper).
Used unsupervised machine learning, natural language processing and google charts API for scientific paper classification (see the demo here).

Postdoctoral Fellow, Canadian Institute for Theoretical Astrophysics (University of Toronto)

Toronto, ON — 2010 - 2014

Developed statistical methods to extract information from large and noisy astrophysical data.

Accomplishments

Made a popular jQuery web-based visualization of sky maps to compare data from various satellite and ground-based astrophysical experiments.
Active member of the data analysis team of the Atacama Cosmology Telescope international collaboration to map the sky in the microwave frequencies and study physical properties of the universe by statistical characterization of those data.
In depth involvement in making a multi-component model of the sky based on large sets of noisy observational data observed at a wide range of wavelengths.
Developed a fast statistical method and wrote the software package in python to precisely cross-calibrate newly observed microwave sky maps using already existing calibrated multi-frequency data over a wide range of frequencies. Maps made from the Atacama Cosmology Telescope data were all calibrated using my code.
Constrained pre-big bang cosmological models using satellite data by searching for concentric circular patterns in 56 high-resolution sky maps (3M pixels each) in three different frequencies. The work was covered overwhelmingly well by the media (see the news coverage of this work here).
Organized and ran Python Lunch Series for Astronomers at University of Toronto (2012-2013).
Taught data analysis at University of Toronto summer schools.
Supervised four students to do research in data analysis in astronomy, research led to publications.

Research Physicist, Department of Astrophysical Sciences (Princeton University)

Princeton, NJ — 2008-2010

Accomplishments

Developed a fast and efficient statistical method for estimating power spectra on small and irregular patches of the sky with minimum loss of information. The method was widely used in the data analysis of the Atacama Cosmology Telescope (ACT).
Taught intermediate algebra in the youth correctional facilities in Princeton, NJ (2009-2010).
Supervised three undergraduate students in the Princeton Astrophysics Undergraduate Summer Research Program (USRP) that led to a publication and a conference presentation.
Served on the National Science Foundation panel to review astrophysical project applications.

Postdoctoral Research Associate, Department of Physics (Princeton University)

Princeton, NJ — 2006-2008

Accomplishments

Developed new statistical methods for novelty detection and pattern recognition in the microwave satellite data.
Made a prediction template for the emission of the free electrons in our galaxy in the microwave frequencies based on the best model for the Milky Way.
Developed a Hamiltonian Monte Carlo algorithm for doing fast and efficient parameter estimation using MCMC methods in cosmology.

SKILLS

Analytical: Problem Solving through Statistics, Data Analysis, Machine Learning, Deep Learning, Neural Networks, Visualization, Probabilistic Programming

Programming: Python, Scala, MATLAB, IDL, Fortran, Java

Analysis: Apache Spark, Theano, scikit-learn, pyMC, Pandas, (No)SQL, NLTK

Development: Flask, HTML5

PUBLICATIONS

I have published more than 60 peer reviewed articles in scientific journals and a number of popular science articles in the newspapers and magazines. A full list of publications can be found on the arXiv (here) or google scholar page (here).

IN THE NEWS

My research in the news and my popular science articles in the newspapers (here).