Ahmed Bahaaeldin Mohammed
Verified Expert in Engineering
Data Scientist and Software Developer
Cairo, Cairo Governorate, Egypt
Toptal member since November 12, 2021
Ahmed is a passionate data scientist interested in solving challenging business problems. He has been working on time series and NLP-related issues for nearly three years at VOIS. He designs novel deep learning architectures that suit the clients' needs, accompanied by statistical inference and modeling. Ahmed helped startups launch MVP products rapidly using state-of-the-art AI models.
Portfolio
Experience
- Generative Pre-trained Transformers (GPT) - 4 years
- Deep Learning - 4 years
- Natural Language Processing (NLP) - 4 years
- Natural Language Understanding (NLU) - 4 years
- Python 3 - 4 years
- Machine Learning - 3 years
- PyTorch - 3 years
- Recommendation Systems - 3 years
Availability
Preferred Environment
Visual Studio Code (VS Code)
The most amazing...
...project I've developed was novel architecture to solve time-series problems combining text input.
Work Experience
Teaching Assistant
Zewail University
- Designed the course flow for the assignments and code review.
- Created Jupyter notebook tutorials to summarize basic NLP skills.
- Explained classical approaches for solving NLP problems.
Applied Scientist II
Microsoft
- Worked on content moderation for Microsoft Bing comments where I developed complex models for the classification of toxicity of each comment.
- Worked on large-scale solutions for thousands of requests per minute.
- Constructed a knowledge base representing the relationship of the topic of the article with the reflected toxicity of people (e.g., to detect what triggers people the most).
ML Engineer
VoiceOps
- Created a novel text-to-text Longformer model that takes calls from call centers and transforms them properly to be shown. Types of ransformations include redaction, punctuation, and diarization.
- Created a "show similar statement" end-to-end pipeline that encoded more than one million events. Users can query most similar events in less than ten seconds.
- Created a clustering pipeline that uses hierarchical clustering combined with dimensionality reduction techniques like (UMAP and PCA) to improve the speed of creating the clusters.
Senior Data Scientist
Vodafone Intelligent Solutions (VOIS)
- Developed a deep learning framework for time series data where the input is a CSV file and the whole training, feature engineering, and preprocessing takes place. Designed multiple deep learning architectures.
- Used the RoBERTa base model finetuned on Stanford NLI dataset for question mapping with FAQ questions in a database with a 74% hit-miss rate. Other trials used BERT, ALBERT, large, and xlarge.
- Improved the ticket routing system by using the XLNet-based model to classify tickets and route them to the resolver group, resulting in an improvement from 60% to 92%.
- Held training for junior and fresh grad data scientists.
Deep Learning Consultant
Arete Global
- Created novel grammar classification architecture with accuracy 82% with ALBERT-base embedded in Conv1D local attention model.
- Developed a factoid question answering system with neural dependency parser and Question recommendation for interviewing bots using the dot product of RoBERTa large embeddings.
- Managed a group of fresh grads and upskilled them in deep learning and NLP.
Experience
Pain Point Detection
https://github.com/ahmedbahaaeldin/Pain-Point-DetectionI mainly used pre-trained transformers and applied multiple experimentations to find the best solution. Moreover, I tried various data augmentation techniques to enrich the dataset and check which augmentation technique improved the overall performance.
Cross-lingual NLP Services
Education
Master's Degree in Machine Learning
Cairo University - Cairo, Egypt
Bachelor's Degree in Computer Engineering
Cairo University - Cairo, Egypt
Certifications
Udacity PyTorch Scholarship
Udacity
Deep Learning Specialization
Coursera
DEV288x: Natural Language Processing (NLP)
Microsoft
Skills
Libraries/APIs
PyTorch, TensorFlow, Pandas, Natural Language Toolkit (NLTK), Torch AI, Scikit-learn, NumPy
Tools
Amazon SageMaker
Languages
Python 3, C++, Python, SQL
Platforms
Jupyter Notebook, Amazon Web Services (AWS), Visual Studio Code (VS Code), Azure
Other
Machine Learning, Deep Learning, Natural Language Processing (NLP), Natural Language Understanding (NLU), Graph Neural Networks, Transformers, Recurrent Neural Networks (RNNs), Convolutional Neural Networks (CNNs), Recommendation Systems, Generative Pre-trained Transformers (GPT), Artificial Intelligence (AI), Computer Vision, Machine Learning Operations (MLOps), Amazon Machine Learning, Data Structures, Algorithms, APIs, Big Data, Generative Adversarial Networks (GANs), Clustering, OpenAI, OpenAI GPT-4 API
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring