Information Security Assistant

OMID ASUDEH

650-***-**** j ****.******@*****.***

Education

Ohio State University Columbus, OH

Master of Science, Computer Science, High Performance Computing and Database, Dec. 2018, GPA:3.53 University of Texas at Arlington Arlington, TX

Master of Science, Computer Science, Information Security, May 2016, GPA:3.87 Qom University of Technology Iran

Bachelor of Science, Computer Science, May 2013, GPA:3.50 Selected Coursework:

Data Mining

Social Networks

Data Analysis

Computer Vision

Advanced Parallel Computing

Advanced Operating Systems

Machine Learning (Stanford University, Andrew Ng, Coursera)

Design and Analysis of Algorithms

Advanced Database Systems

Advanced Information Security

Secured Programming

Skills

Web Programming: ASP.NET, Razor, MVC, PHP, HTML, JavaScript, JQuery, CSS Programming Languages: C++, C#, Python, Java, C

Parallel Programming: Pthread, OpenMP, MPI (Message passing), CUDA (GPU) Teaching: TA for Intro. to Database Systems, Fall 2017, Spring 2017, Fall 2016, Ohio State University. Other: Git, Azure Devops, Data Analysis, SQL, Matlab, OpenCV, experienced with large code debug, and TDD.

Experience

Software Engineer - UX Microsoft, WA, Bellevue, Sep 2019- Nov 2020 World-Class UX development for Microsoft Bing Ads using technologies such as MVC, C#, ASP.Net, Azure DevOps and multiple test driven development. The tasks included but were not limited to creating new mocks, data-gathering, development/testing of the ideas, experimenting ights using Bing.com real tra c and human judgements, statistical analysis of the results, and nally shipping of the successful experiments.

Jr. Data Scientist Tcetra, Ohio, 2019 Mar-Sep

I was responsible for daily data cleaning using ETL and SSMS and also Python libraries like Pandas, providing summary reports and designing the DataCube + collaboration with UX developers Graduate Research Assistant HPC Lab, Ohio State University, 2016-2018

*Designed, developed and tested a high performance Approximate Aggregation Matrix Querying interface.

The tool was aimed to quickly answer the

aggregate queries on Big Data with a high

accuracy.

We utilized the bit vectors to simulate a

search tree data structure built on the top

of preprocessed aggregate data.

The tool achieved performance improvement

with minimized space overheads.

MPI, OpenMP, Data Analytics

*Designed, implemented, and tested a Parallel Bitwise Operation toolkit that works on huge sized compressed bit vectors.

We developed a fair load balancing ap-

proach to distribute the work between mul-

tiple nodes and threads.

C++, MPI, pthread, OpenMP, CUDA

Graduate Research Assistant ISEC Lab, University of Texas at Arlington 2014-2016

Implemented a real-time website phishing detection tool (a browser add-on) that detects/alerts the malicious websites based on their visual imitation of the actual websites. The research was published in 2016 ACM SIGSAC. The tool minimizes the false positives/negatives using a multi-phase shallow and deep detection approach that bene ts from machine learning models trained on a huge-sized datasets crawled from the web. (Python, Machine Learning, Computer Vision, PHP, Javascript, OpenCV, Matlab, Data Analytics).

Designed, developed and tested a decision making/recommendation tool to sys- tematically suggest the best tted authentication scheme to the decision makers based on their requirement. (PHP, Javascript, Machine Learning, Recommender Systems). University Projects

Approximate Aggregation over Spatial Data using Bitmap Designed and developed algorithms/techniques to answer approximate aggregate queries over huge matrices with high performance. Our tool was able to answer the huge aggregate queries over matrices 10X faster than the baseline approach. The idea was to represent the ag- gregate data as a search tree and query that instead. The issue is that the tree gets huge in terms of space, bigger that actual data. Therefore, the approach simulate the search tree with bitmap indexes that occupies much less space. We also utilized multiple work distribu- tion/parallelization techniques for the implementation side. (C++, MPI, OpenMP, pthread, CUDA, Data Analytics).

Parallel Bitwise Operations on Word-aligned-hybrid compressed bit vectors Implemented a toolkit for Parallel Bitwise Operations on Word Aligned Hybrid (WAH) Com- pressed bit vectors without decompressing them. The proposed approach achieved 16X per- formance improvement for bitwise AND and bitwise OR. The main challenge there was to perform a fair load balancing technique considering the loss of random access due to the compression. (C++, MPI, pthread, OpenMP, CUDA).

Multi-phase Real-time Website Phishing Detection using Visual Similarity Designed and Implemented a method to detect the Phishing websites in a real-time manner with minimized false negatives based on the fact that they are visually masquerading the famous websites. The work was published in ACM SIGSAC 2016. We developed a browser add-on based that could detect fake website with a high accuracy and in a real-time fashion. Each detection phase has its method to detect the phishing websites from shallow visual detection to the deep trained models. (Python, Machine Learning, Computer Vision, PHP, Javascript, OpenCV, Matlab, Data Analytics).

Authentication Scheme Suggesting Tool Designed and developed a Web-based recom- mendation system that suggests a best t Authentication Scheme to the decision makers based on their needs using decision trees, trained models, and nearest neighbor classi cation models. In a survey we have done, users indicated that the generated recommendations by the tool can potentially save a lot of money and time for the companies by avoiding expensive switching between multiple authentication approaches. (PHP, Javascript, Machine Learning, Recommender Systems).

Real-time Pedestrian Detection, tracking, and Distance Estimation Designed and implemented algorithms to detect and track the pedestrians and estimate their distance from the camera in a video stream in a real-time manner. Our model was able to successfully detect and track the pedestrians with a high accuracy in a real-time fashion. We used a combination of detection and tracking algorithms (HOG, Lukas Kanade) and also trained models for our purpose. We developed new heuristics for distance estimation and 3D comprehension without Kinnect. (C++, OpenCV, Computer Vision) Publication and Awards

O Asudeh, M Wright. \Phishing Website Detection with a Multiphase Framework to Find Visual Similarity". Proceedings of the 2016 ACM SIGSAC.

GRA/GTA fellowship, The Ohio State University 2016 - 2018

GRA fellowship 2015, UT Arlington.

Grad. Scholarship in Theory Based Web Systems 2014, UT Arlington.

Contact this candidate