Hi, I'm Selssabil Boudefel

AI & Data Science

A Computer Science student specializing in Artificial Intelligence and Data Science, passionate about Machine Learning and Deep Learning with a growing interest in Reinforcement Learning and also AWS servcies. I get the job done, learn fast, adapt easily, and love solving problems. Always curious, always building.

About Me

Computer Science Student

Currently, I'm working on my final year project focused on recommendation systems. Once I complete it, I'm excited to start my career in Machine Learning or Deep Learning and help businesses make smarter, better, and data-driven decisions.

My Projects

PROJECTS & Work Experience

Kinship Verification

Developed a computer vision project specializing in kinship verification to accurately detect relationships between individuals.

Utilized OpenCV for robust image processing, extracting essential features required for accurate kinship prediction.

Incorporated facial descriptors including LPQ, BSIF, and HOG to extract vital features for reliable kinship analysis.

Leveraged widely recognized datasets, including the Cornell, KinFace I and II, to train and evaluate the model, ensuring real-world applicability.

Churn Prediction using SageMaker

Executed a comprehensive data science project, leveraging AWS SageMaker for proficient model training, with a primary focus on customer churn predictions.

Covered the full project lifecycle within a single notebook, from initial data exploration and visualization to rigorous data preprocessing, model training, and deployment.

Computer Vision Researcher Intern

Contributed as a Computer Vision Researcher Intern, leading a Kinship Verification project with successful outcomes.

Enhanced my communication and organizational skills through valuable experiences gained during my internship.

Amplified my proficiency in exploring, analyzing, and manipulating datasets, strengthening my data analysis skills.

-

Stock Price Prediction

Developed and implemented trading models by leveraging Tesla ‘s historical trade data

Utilizing machine learning and Time Series techniques to optimize trading strategies and enhance predictive results.

Text To Audio Solution App

Developed a text-to-audio transformation application using Python (Tkinter) for the front end.

Leveraged AWS services, including AWS Lambda, Polly, S3, SES, and Parameter Store for seamless backend functionality.

Leveraged AWS services, including AWS Lambda, Polly, S3, SES, and Parameter Store for seamless backend functionality.

The Cloud Resume Challenge - AWS

Successfully completed The Cloud Resume Challenge by building and deploying a personal resume website using Amazon Web Services.

Utilized AWS services, including Amazon S3, AWS Route 53, and AWS CloudFront, to securely host my website.

Implementing GitHub Actions to automate updates to the associated Amazon S3 bucket upon each code push.

Integrated a visitor counter using AWS DynamoDB, AWS Gateway, and Lambda function to capture and update visitor statistics dynamically.

My Blogs

Exploring SwiGLU : The Activation Function Powering Modern LLMs

This blog explores SwiGLU, an advanced activation function that enhances performance in modern large language models (LLMs). It looks at SwiGLU's math, its benefits over traditional functions, and how to implement it using PyTorch