Data Science Fellow

MATINEH SHAKER

319-***-**** ******.*******@*****.***

https://www.linkedin.com/in/matinehshaker

https://github.com/matinehshaker

EXPERIENCE

Insight Data Science Sep 2015 - Present

Fellow New York City, NY

Product Predicted monthly returns of hedge funds with short history as a consultant for Sliced Investing, with blog post available at http://matinehshaker.github.io/hedgeclone/

Algorithm Built a python-based tool that uses Bayesian regression through Gaussian processes to compute the factor loadings of a hedge fund on di erent risk factors given its monthly returns.

Result The prediction results improved the previous method by 20 percent, and also provided con dence intervals. The factors with highest estimated weights were validated by the company to be the important factors in that fund based on domain knowledge.

Brigham and Womens hospital, Harvard Medical Jan 2015 - May 2015 Predoctoral Fellow Boston, MA

Anomaly detection: Applying Gaussian graphical models for maximum likelihood estimation of a normal model of brain di usion signal using graphical Lasso algorithm. The model is used for anomaly detection in traumatic brain injuries. As a result detection rate improved by 20 percent, while model complexity reduced by 10 times. Northeastern University Jul 2012 - Present

Research Assistant Boston, MA

Machine learning: Developed a novel algorithm for nonlinear manifold learning based on density distribution of data using principal surfaces. The tool is used for visualization and classi cation of high dimensional data.

Graph algorithms: Designed an e cient algorithm design for calculation of kernel density derivative estimate with 3 times reduction in number of operations.

Statistical signal processing: Built maximum likelihood matched lter for detection and estimation of optical intravascular signals, which improved detection of low-amplitude targets by 18 percent. Eigen Jun 2011 - May 2012

R&D Engineer Grass Valley, CA

Computer Vision: Developed mutual information image registration algorithm for a prostate biopsy device to combine information from ultrasound and MRI imaging devices. SKILLS

Computer Languages Python, R, MATLAB, Familiar with C++ Tools Scikit-Learn, Pandas, ITK, Git, SVN, Latex, Familiar with MySQL, Flask EDUCATION

Northeastern University Jul 2012 - Present

PhD in Electrical & Computer Engineering

University of Iowa Dec 2011

MS in Computer Engineering

University of Tehran Aug 2008

BS in Electrical & Biomedical Engineering

AWARDS

Travel award for IPAM workshop UCLA (2013) College of Engineering scholarship, University of Iowa (2010)

Best paper award at imaging informatics conference, Iran (2008) Ranked top 0.3 percent in national university entrance exam, Iran (2004) National mathematics Olympiads semi nalist, Iran (2003)

Contact this candidate