Open to opportunities

Marwa Matar

@marwamatar

AI/ML researcher and engineer specializing in NLP, computer vision, and speech—building and deploying performant deep learning systems.

Egypt

Message

What I'm looking for

I’m seeking a challenging role where I can leverage my expertise in machine learning, deep learning, computer vision, data analysis, data science, NLP, and deployment (REST APIs, Docker) to drive impactful results.

I’m a highly motivated AI/ML professional with extensive experience in research and development. I specialize in machine learning, deep learning, natural language processing (NLP), image processing, speech processing, and software engineering.

On the Stars team at Qur’an QA (Qatar University), I contributed to an extractive question answering system for the Holy Qur’an and achieved top performance on the Qur'an QA 2022 shared task. I also developed and released a new dataset, built a QA model based on ArabicBERT, improved it with K-fold cross-validation and ensemble methods, and published a paper.

As a Research Assistant at National University of Ireland, Galway, I developed ML and DL models for NLP tasks (including suggestion mining) for the PANDEM-2 project. Earlier, during an internship and freelance role at Fusionbase München, I built data pipelines with Python and extracted, cleaned, and transformed unstructured text into useful datasets.

My background also includes embedded software engineering at Bio-Business Giza, where I developed embedded software using C and MATLAB, configured and tested a Microcontroller Abstraction Layer, and tuned and tested PID control. I’m currently a PhD Researcher at Al-Azhar University, building multimodal emotion recognition systems and deploying AI models as REST APIs in reproducible Dockerized pipelines.

Experience

Work history, roles, and key accomplishments

Participant (Stars team)

Qur’an QA, Qatar University

Feb 2022 - Jun 2022 (4 months)

Contributed to an extractive question answering system for the Holy Qur’an, achieving top performance on the Qur’an QA 2022 shared task. Developed a new dataset and enhanced an ArabicBERT-based model using K-fold cross-validation and ensemble methods, and published a paper.

NLP Extractive Question Answering Transformer Models Datasets Cross Validation Machine Learning (ML)

Research Assistant

National University of Ireland, Galway

Jul 2021 - Nov 2021 (4 months)

Developed machine learning and deep learning models for NLP tasks, including suggestion mining, for the PANDEM-2 pandemic surveillance project. Designed and implemented software components for preprocessing, modeling, and prediction using social media data and standard datasets.

Machine Learning (ML)Deep Learning Suggestion Mining Modeling Prediction NLTK WordNet SentiWordNet SpaCy NLP

Data Analyst Internship and Freelancer

Fusionbase München

Oct 2020 - Apr 2021 (6 months)

Built reliable data pipelines in Python to transfer data from the web to Fusionbase infrastructure. Extracted, cleaned, and transformed unstructured text from multiple sources into useful datasets to improve data quality.

Python Data Pipelines Data Extraction Data Cleaning data transformation

Embedded Software Engineer

Bio-Business Giza

Jun 2012 - Jun 2014 (2 years)

Developed embedded software using C and MATLAB for projects including incubators, sperm analyzers, and patient monitors. Developed and tested the ECUAL microcontroller abstraction layer (I/O, keypad, interrupts, ADC, UART, I2C), troubleshot embedded systems with lab instruments, and tuned PID.

Function Generators Embedded Software C#Matlab Microcontroller Abstraction Layer (ECUAL)UART I2C Interrupts Keypad I O Function Generator

Education

Degrees, certifications, and relevant coursework

Al-Azhar University

Doctor of Philosophy (PhD), Computer Vision

2023 -

Activities and societies: Minor: Computer Vision. Containerizes applications with Docker for reproducibility and portability.

Conducts research on multimodal emotion recognition in e-learning using deep learning, focusing on facial features. Builds inference pipelines combining face detection (OpenCV, Dlib) with CNN-based emotion classification and deploys models as REST APIs.

Al-Azhar University

Master of Science, Natural Language Processing

2015 - 2019

Activities and societies: Minor: Natural Language Processing (Arabic NER). Addressed Arabic tokenization and morphological challenges.

Pursued an MSc focused on AI for Arabic Named Entity Recognition. Implemented LSTM and CRF-based models and performed data preprocessing, feature engineering, and model evaluation for Arabic text corpora.