- Utilized Python programming language and Bash scripting to preprocess and extract features from 4D complex brain imaging data amounting to 0.25 TB, applying a 3D multiresolution optical flow technique.
- Employed signal processing techniques with Python programming language to modulate respiratory signals with calculated brain signal speed.
- Conducted A/B testing to identify significant differences in the modulation of brain cardiovascular pulse with respiration between control subjects and individuals with Alzheimer's disease.
- Visualized and presented research findings to neurological researchers, contributing to a greater understanding of Alzheimer's disease. Additionally, authored and submitted a paper to JCBFM.
- Paper: Elabasy, A., Suhonen, M., Rajna, Z. et al. Respiratory brain impulse propagation in focal epilepsy. Sci Rep 13, 5222 (2023). https://doi.org/10.1038/s41598-023-32271-7
- Trained multiple classifier that improved the classification performance of imbalanced Alzheimer’s fMRI images by more than 12% compared to state of art on the same data.
- Analyzed, visualized, and discussed the results with a team of neurological researchers to have a better understanding of the results and Alzheimer’s disease.
- Analyzed, visualized, and reported the results and submitted a research paper to ISPr 2023 scientific conference.
- Developed a real-time sign language interpretation application using React.js, tensroflow and tensorflow.js and deployed on IBM cloud servies.
- Building a stable diffusion web application using Hugging Face, React, and deployed on fastAPI.
- Building a recommendation engine using Alternating Least Squares in PySpark and using the popular MovieLens dataset and the Million Songs dataset.
- Building a real time car plate detection mobile application using Tensorflow and EasyOCR.
- Automobile price prediction: Utlitize python to implement end to end data science pipeline to predict the price of old Automobile based on the given features.
- Sensor Activity Recogniation: Classifying the output of eight sensors into five activities and studied the effect of changing window sizes and axel combination.
- Alzhimers CV-BOLD Classification: Utilized Python to develop supervised machine learning techniques to classify imbalanced Alzheimer’s CVBOLD data, which enhanced the classification performance by 10%.
- Find the best location to open a new Gym: Utilized python to implement unsupervised techniques to helping the business owner to increase his revenue by finding the best neighborhood to open a new gym.
- Customer identification for mail order products: Utilized python to implement unsupervised techniques to helping the business owner to increase his revenue by finding the best neighborhood to open a new gym.
- Melenoma Classification: Classifying malignant Melanoma using skin lesion images using CNN-based classifiers.
-
Object Tracking using Particle Filter: Implemented particle filter to track walking object in video
-
Pose Estimation and Squat counter: Utilize python to develop a real-time pose estimation and squat counter using MovingNet lightning.
- Sentiment Analysis web app: Web application for classification of reviews, using deep learning model implemented in PyTorch and deployed on Amazon SageMaker.
- Plagirasm Detector web app: Creating plagiarism detector trained on LSC and containments features and deployed on AWS SageMaker.
- Data Science Resume Selector: Selecting the resume that are eligbile to data scientist postions, the dataset used contains 125 resumes, in the resumetext column. Resumes were queried from Indeed.
- Power consumption prediction: Real time prediction for power consumption using DeepAR on AWS.
- Immigrants to Canada data visulization: Visualizing the data of the immigrants to Canada using different visualizing libraries in Python.
- Geospatial visualization of San Francisco Police Department Incidents: Visualizaing the geospatial data of the San Francisco police department incidents for the year 2016.
- San Diego Rainforest Fire Predicition: Predicting the occurance of rainforest fire in san Diego using weather data collected by san Diego weather center.
- Cluster Analysis of the San Diego Weather Data: Ultilizing pyspark to implement unsupervised learning model to cluster the san Diego weather data so as to better understand the occurance of the rainforest fire.
- Songs App User Activity Data Modeling : Modeling user activity data for a music streaming app called Sparkify to optimize queries for understanding what songs users are listening to by creating a Postgres relational database and ETL pipeline to build up Fact and Dimension tables and insert data into new tables.
- Songs App data modeling using Apache Casandra: Create an Apache Cassandra database which can create queries on song play data to answer analysis questions.
- Machine Learning Specialization
- Big Data Specialization
- Intro to Machine Learning with Tensorflow Nanodegree
- Machine Learning Engineer Nanodegree
- AWS Fundemntals Specialization
- Data Science Professional Certificate
- AI for Medicine
- Data Visualization with Seborn
- Time Series Forecasting
- Data-Visualization-with-Plotly-in-Python
- Machine Learning Explanality-Kaggle