Abdulrahman Adel

Machine Learning Engineer

omaline•  January 2024 - Present•  Remote

As a Machine Learning Engineer, I lead initiatives to enhance user experience and engagement through innovative technology and strategic collaboration. Highlights of my role include: * Developing a dynamic search engine using Large Language Models (LLMs), significantly improving relevance. * Engineering algorithms to optimize search result ranking, boosting user engagement. * Orchestrating large-scale web scraping for data acquisition and model training. * Fine-tuning LLMs for efficient processing of datasets. * Building machine learning models for product categorization, enhancing search accuracy. * Collaborating cross-functionally to align improvements with user needs. * Conducting rigorous testing to ensure performance and scalability. * Contributing to technical strategy for enhancing search capabilities.

Research And Development Engineer

Iskraemco•  October 2021 - June 2022•  Cairo, Egypt

* Assisted in hardware design of Next Generation Smart-Grid Meter (NSGM). * Developed C++ script for efficient data collection and transmission to cloud. * Implemented PyTorch-based deep learning model for power consumption forecasting. * Utilized variational autoencoder for non-intrusive load monitoring. * Designed algorithm for harmonics and anomaly detection, enhancing grid stability.

Machine Learning Engineer

Aprcot•  August 2021 - October 2021•  Cairo, Egypt

* Developed Arabic Automatic Speech Recognition (ASR) prototype. * Designed Transformer-based encoder-decoder architecture. * Achieved 12% Word Error Rate (WER) on Mozilla Common Voice dataset. * Utilized Google Cloud Platform (GCP) with GPU for efficient model training. * Collaborated with development team for chatbot integration.

Machine Learning Researcher

WRL - Wireless Research Lab•  September 2020 - July 2021•  Cairo, Egypt

* Developed AI-based module to detect cheating in online lab exams using mouse interaction analysis. * Employed KNN, SVC, Random Forest, Logistic Regression, XGBoost, and LightGBM algorithms for classification. * Conducted experiments validating effectiveness of approach with up to 90% accuracy using LightGBM. * Achieved 88% precision and 95% degree of separation in cheat detection.

Helwan University

Computer Science & Engineering, BE•  September 2017 - June 2022

Note: Good