Blog - منصة قلم التعليمية

0.00 (0.00)

Amira Fouda

0 دورة • 0 طالب

0.00 (0.00)

سيرة شخصية

Profile

A scientific researcher in artificial intelligence, multimedia, and data visualization field.

Interested in machine learning, computer vision and specifically natural language processing. Seeking to fill a position as an AI engineer.

Education

9-Month Professional Training Program Information

Technology Institute (ITI) Artificial Intelligence (AI) Track

Graduation project: A-Eye Image Caption (NLP&CV)

With Grade: Excellent

Bachelor of Computer Science, Cumulative grade:

Excellence, with honors, ranked 2nd, Obour Institutes

Graduation Project with distinction:Twitter Sentiment

Analysis Natural Languag Processing (NLP)

Projects

RAG-Link

Developed a Retrieval-Augmented Generation (RAG) system utilizing the latest LLMs like Mistral 7B, Llama 3, GPT-4-turbo, and Gemini Pro for enhanced text generation and retrieval.
Deployed Ollama locally to fine-tune and run models efficiently for optimized RAG performance.
Created high-quality embeddings using BGE, E5, and OpenAI text-embedding models.
Implemented ChromaDB as the vector store for efficient retrieval across multiple document types including webpages, PDFs, CSVs, Excel files, and databases.
Designed the system to handle question-answering, report generation, and data visualization for actionable insights.

Retail Action Recognition

Designed and implemented a computer vision-based system for analyzing retail environments.
Developed AI models capable of detecting persons and objects within retail stores.
Integrated real-time tracking of both customers and retail items to enhance security and store analytics.
Built human action recognition models to analyze customer behavior and detect potential theft or unusual activities.
Leveraged Deep Learning (YOLO, OpenCV, and Transformer-based models) for robust action recognition and anomaly detection.

Validation System

Developed a document verification and validation system using cutting-edge OCR and Biometric Authentication technologies.
Integrated AI-powered OCR to extract and validate text from various document formats such as passports, identity cards, invoices, and official papers.
Implemented facial recognition and fingerprint authentication to enhance security and prevent fraudulent access.
Optimized the system for real-time processing, ensuring efficient verification for businesses and secure user authentication.
Designed a scalable and modular architecture to support multiple document types and authentication methods

LPG Vending Machine

Designed an automated validation and detection system for LPG vending machines.
Implemented YOLOv8-based computer vision models to detect and validate the presence of LPG cylinders.
Developed classification models to identify different cylinder types and their respective statuses.
Integrated real-time slot detection to monitor available and occupied cylinder slots.
Optimized the system for edge computing to ensure fast and reliable processing on vending machines.
Enhanced the security and efficiency of LPG vending operations by providing automated monitoring and validation.

Arabic ASR (Automatic Speech Recognition)

Developed an Arabic Speech-to-Text (ASR) system to transcribe spoken Arabic into written text accurately.
Collected and preprocessed Khaligi and Arabic speech data using web scraping techniques from YouTube.
Utilized Wav2Vec2 and pre-trained large language models (LLMs) from Hugging Face to improve transcription accuracy.
Implemented real-time voice recognition and text autocompletion for enhanced user interaction.
Optimized model performance for various Arabic dialects by fine-tuning on diverse datasets.

Sign Language Recognition

Developed a bidirectional Sign Language Recognition system to bridge communication gaps for the hearing-impaired community.
Implemented computer vision and deep learning models to convert sign language gestures into spoken language in real-time.
Integrated speech-to-sign translation, converting spoken language into corresponding sign language animations.
Collected and preprocessed large-scale sign language datasets to improve recognition accuracy.
Utilized transformer-based and CNN-based models (e.g., MediaPipe, OpenPose, and LSTMs) for hand and gesture detection. Designed an intuitive UI for real-time sign language interpretation, making communication more accessible.

Customer Service Chatbot

Developed an AI-driven Customer Service Chatbot for automated support and improved customer engagement.
Utilized GPT-3.5 and large language models (LLMs) for advanced intent recognition and response generation.
Integrated pretrained NLP models from Hugging Face to enhance conversational understanding.
Designed a context-aware chatbot capable of handling multiple conversation threads efficiently.
Implemented multi-turn dialogue processing and personalized responses to improve user satisfaction.
Optimized chatbot performance for scalability and deployment across multiple platforms.

Real-time Head Pose Estimation

Supervised machine learning - Linear Regression -Media pipe - DLib

Detective Fake Jobs | Kaggle competition,

Deep learning - natural language processing (NLP)

State Farm Distracted Driver Detection

Deep learning - computer vision (CV)

NewsClassification

Using LSTM - NLP

Dry Beans Classification | Kaggle competition

Supervised Learning - Classification

Identify Wine

Unsupervised machine learning - Clustering - Kmeans -DBscan - Hierarchical - GMM -PCA

Wuzzuf Jobs Analysis

Dealing with data by spark - Java - Using Spring Boot to deal with a web server.

Skills

Programming Languages : C | Cpp | Python

MLOps Skills : Docker | Flask | AWS | FAST API

Soft Skills

Leadership
Team Player
Organization
Problem solving
Stress management
Creativity
Time management

الدورات

لا توجد دورة حتى الآن.