Mory Gharasuie

Email: ad2id9@r.postjobfree.com Phone: +1-757-***-****

GitHub: https://github.com/mortezamg63

LinkedIn: https://linkedin.com/in/morteza-mohammady-53415258 Summary

I am a PhD candidate at Old Dominion University (ODU) with research topics on machine learning in semi-supervised, self-supervised and unsupervised learning on image and structured

(tabular) data, plus experience in computer vision and familiarity with NLP+LLMs. Currently, seeking an intern role in Machine Learning (ML), Computer Vision or Data Scientist fields. Skills

Programming language: python C/C++ C# Java SQL MATLAB Big Data: Apache Spark Hive

Libraries: TensorFlow (Graph, API, and Eager modes) PyTorch Torchvision OpenCV Scikit-learn Whoosh NLTK Hugging Face Seaborn imblearn Scipy pandas Research skills: Machine Learning Neural Networks Deep Learning Self-Supervised Learning

Semi-Supervised Learning Tabular Data Computer Vision Image Processing Natural Language Processing Transformers Autoencoders GANs Others: Jupyter Notebook Git Docker

Education

PhD in Computer Science August 2019-present

Old Dominion University Norfolk, USA Expected Graduation: July 2025 MSc in Artificial intelligence August 2011-July 2013 University of Nabi Akram Tabriz, Iran

BSc in Computer Engineering Aug 2008

University of Shahid Shamsipour Tehran, Iran

Academic Experience

Publications

● Gharasuie, M. M., & Seyedarabi, H. (2013, September). Real-time dynamic hand gesture recognition using hidden Markov models. In 2013 8th Iranian Conference on Machine Vision and Image Processing (MVIP) (pp. 194-199). IEEE.

● Gharasuie, M. M., & Gaffari, A. (2015, November). An efficient run-based method for connected component labeling. In 2015 9th Iranian Conference on Machine Vision and Image Processing (MVIP) (pp. 100-104). IEEE.

● Gharasuie, M. M., Jennings, N., & Jain, S. (2021, June). Performance monitoring for exercise movements using mobile cameras. In Proceedings of the Workshop on Body-Centric Computing Systems (pp. 1-6).

● Gharasuie, M. M., & Wang, F., (2022). Progressive Feature Upgrade in Semi-supervised Learning on Tabular Domain. IEEE International Conference on Knowledge Graph (ICKG)

● Gharasuie, M. M., Sharif, O., & Wang, F. “SAWTab: smoothed Adaptive Weighting for Tabular Data in Semi-Supervised Learning”, PAKDD 2024. Research Experience

● Target-based encoding for semi-supervised learning on tabular data: In this study I proposed a target encoding non-numerical data in semi-supervised settings. Unlike current one-hot encoding, my approach suits large datasets with high cardinality categorical features. I've published one paper, with another under review.

● BERT Model and its variations: Applied tabular data to the BERT model or its variations and studying their performance on various datasets. In this project, I used different popular preprocessing techniques in BERT. The results show accuracy improvement on some datasets .

● Current project: I am working on semi-supervised learning with focus on addressing bias in imbalanced datasets, applicable to both images and tabular datasets. The goal is to tackle bias arising from uneven sample distributions in multi-class and binary classification scenarios.

Educational Experience

● Information retrieval course: NIH estimates 80% clinical trials miss recruitment timelines, and many fail to enroll minimum participants. We designed a retrieval system for finding eligible trials in a dataset from ClinicalTrials.gov. I contributed to data parsing, selection, ranking, and coding. Our approach used two ranking steps: BM25F and BlueBERT for embeddings via contrastive learning. It improved over just using one ranking step by 13%.

● Natural Language Processing course: Studied recent architectures in LP domain and different tools and concepts for processing natural language (known as text-to-code). The goal of the course project was to design a program to build an HTML web page based on natural language for non-expert users. For implementation, I used name entity recognition and NLTK library to process the input natural language, and BeautifulSoup library for generating the HTML.

● Data Science and Big Data course: Studied concepts in Distributed computing and working on projects related to big data processing in Apache Hadoop, Apache Spark, Hive, and Scala.

● Mobile Sensing for Smart Cities course: applied a study on video analytics, aiming to reduce the resource requirements for vehicle detection in a surveillance camera. I explored diverse configurations including object detection models (YOLO, Fast R-CNN and so on), frame rates, and frame sizes, to dynamically optimize accuracy and resources in the existing video analytic pipeline. I developed the pipeline in PyTorch. Work Experience

Old Dominion University Norfolk, VA, USA 2020-Present Computer Science Teaching Assistant

● Programming with C++ (CS150, CS250)

● Programming with Java (CS151, CS251)

Technical and Vocational University Qom, Iran September 2014-July 2019 Lecturer

● Web Application Development (C#, ASP.NET)

● Data Structures and Algorithms (C/C++ and Pascal)

● Operating System Concepts

● Operating System Lab (Linux)

● Supervised 50 software projects from inception to deployment (Java, PHP, Unity) Computer Networking Executive Qom, Iran January 2014-July 2019 Several companies, firms, and organizations

● Network: Wireless communication among different branches of Mehr Bank in Qom, Iran as MAN Wireless engineer. My responsibility was to add and maintain all their wireless communications, ensuring they are up and running. Also, the same services had been provided to other companies, firms, and educational centers.

● Fibre Optic: installed more than 25000 out of 75000 meters of FTTC and FTTH in a city-wide scale, Qom, Iran.

Programming Manager Qom, Iran January 2013-July 2019 Royan Company

1. PHP

● Designed and customized websites using Joomla and WordPress for customers for small businesses.

● Created web-based administration interfaces for Linux application servers (such as chat server, FreeRadius server, and Elastix) that can be utilized across various operating systems. Enhancing the panels with support for multiple languages and designing them to be more intuitive and user-friendly, tailored to meet the specific needs and preferences of the customers.

2. ASP.NET and C#: Developed websites for small and medium-sized enterprises. Software Developer Tehran, Iran June 2010-August 2011 Pazhang Corporation

● A web-based application was developed to facilitate paperless office operations. The application was specifically designed to streamline the process of forwarding internal letters among various offices and individuals within a hierarchical structure, enabling managers and their respective teams across the country of Iran to access and collaborate on the platform.

Contact this candidate