Anas Tarek

Artificial Intelligence Intern

Aug 2024 - Sep 2024 · 2 mos

Telecom Egypt · Internship · Alexandria, Egypt · On-site

Bachelor's Degree in Software Industry & Multimedia

Present

Faculty of Science, Alexandria University

I'm a student at the Faculty of Science in the Software Industry and Multimedia department in Alexandria University | CGPA (3.5/4)

Cloud Resume Challenge

June 2025 - Present

Technologies: AWS (S3, CloudFront, Lambda, API Gateway, DynamoDB, Route 53), Terraform, GitHub Actions, HTML, CSS, JavaScript

• Developed and deployed a static resume website built with HTML, CSS, and JavaScript.
• Hosted the static site on AWS S3 and served it globally with low latency using AWS CloudFront as a CDN.
• Configured Amazon Route 53 to manage a custom domain, directing DNS queries to the CloudFront distribution.
• Built a serverless backend using AWS Lambda and API Gateway to handle a visitor counter feature.
• Utilized Amazon DynamoDB to store and retrieve the website's visitor count.
• Automated the entire infrastructure provisioning process using Terraform for Infrastructure as Code (IaC).
• Implemented a CI/CD pipeline using GitHub Actions to automatically deploy changes to the frontend and backend upon code commits.

Soul Support

2024-2025

Technologies: Flutter, Dart, Python, TensorFlow, Keras, Scikit-learn, NumPy, JSON, Natural Language Processing (NLP)

• Front-End Development: Architected and developed the entire patient registration system from the ground up using Flutter. This involved creating a seamless, user-friendly interface for patient onboarding, ensuring a smooth and intuitive user experience.
• Chatbot Integration: Designed and developed "Soul Mate," a conversational AI to serve as a personal therapeutic assistant on the Soul Support platform.
• AI & Machine Learning: Sourced and structured a comprehensive dataset of intents, patterns, and responses in a JSON file to define the chatbot's knowledge base on topics ranging from emotional states (sadness, anxiety) to factual mental health information.
• Built, trained, and deployed a deep learning model for Natural Language Understanding using Python with TensorFlow and Keras libraries.
• Engineered a data preprocessing pipeline to vectorize text data using the Keras Tokenizer and encode labels with Scikit-learn.
• Implemented and trained a sequential neural network with an Embedding layer, achieving high accuracy in classifying user intent after 550 epochs of training.
• Developed a Python-based interactive interface that uses the saved model to predict user intent and generate contextually appropriate, empathetic responses.

AWS Multi-Tier VPC Architecture

August 2024

Technologies: AWS VPC, Subnets, Internet Gateway, NAT Gateway, Route Tables

• VPC Creation: Created a custom VPC with a specified IPv4 CIDR block (e.g., 10.0.0.0/16).
• Subnet Creation: Designed and implemented both public and private subnets across different availability zones to ensure high availability and fault tolerance.
• Internet Gateway: Attached an Internet Gateway to the VPC to enable internet access for the public subnet.
• NAT Gateway: Deployed a NAT Gateway in the public subnet to allow instances in the private subnet to securely access the internet.
• Route Tables: Configured route tables for public and private subnets to control the flow of traffic.

Semantic Segmentation with Amazon SageMaker

2024

Technologies: Amazon SageMaker, S3, IIIT-Oxford Pets Dataset, Machine Learning

• Data Preparation: Downloaded and organized the IIIT-Oxford Pets Dataset for use with SageMaker's semantic segmentation algorithm. Prepared the data with the correct folder structure, splitting it into training and validation sets.
• Notebook Instance: Created and configured a SageMaker notebook instance to handle model training and deployment.
• Data Upload to S3: Set up an S3 bucket and uploaded the dataset for access by the SageMaker environment.
• Estimator and Hyperparameters: Created a SageMaker estimator, specifying training instances and setting up hyperparameters for semantic segmentation.
• Model Training and Deployment: Trained a semantic segmentation model using SageMaker's built-in algorithms. Deployed the model to an endpoint for real-time inference.
• Inference and Endpoint Management: Conducted inference with the deployed model and cleaned up resources by deleting the endpoint post-inference.