Mohamed A. Abdelhady, Developer in Amsterdam, Netherlands
Mohamed is available for hire
Hire Mohamed

Mohamed A. Abdelhady

Verified Expert  in Engineering

Bio

As a machine learning tech lead, Mo has successfully taken multiple products and teams from 0 to 1. He uses Python to build ML products that analyze billions of transactions daily. Mo has over eight years of experience working with companies all around the world at different scales. When he's not training ML models, Mo enjoys traveling, hiking, and cooking. Mo also enjoys learning new things. He is currently learning about web development.

Portfolio

Adyen
Artificial Intelligence (AI), Data Science, Deployment, Scikit-learn, PySpark...
Fourthline
Computer Vision, Python, TensorFlow, Large Language Models (LLMs)
ASML Holding
Python, System Design, Modeling

Experience

  • Git - 7 years
  • Data Analysis - 5 years
  • Python - 5 years
  • TensorFlow - 5 years
  • Machine Learning - 3 years
  • Data Science - 3 years
  • Large Language Models (LLMs) - 2 years
  • Retrieval-augmented Generation (RAG) - 2 years

Availability

Part-time

Preferred Environment

Python, Git, Pandas, FastAPI, Docker, Artificial Intelligence (AI), Computer Vision, Data Science, Deep Learning, Vercel

The most amazing...

...thing I've done recently is develop a graph-based deep learning model that flags fraudulent users and saves over €1.3 million for our clients.

Work Experience

Machine Learning Tech Lead

2021 - PRESENT
Adyen
  • Spearheaded the research and development of machine learning-driven products for external and internal fraud detection use cases and anti-money laundering (AML) applications.
  • Developed and deployed graph-based machine learning models, leading to a significant increase in the recall of fraud detection products and help saving platforms more than $2.5 million.
  • Grew and upskilled the team and introduced workflows to ensure data governance and software development best practices.
Technologies: Artificial Intelligence (AI), Data Science, Deployment, Scikit-learn, PySpark, Hadoop, Jupyter, Pytest, Git, Variational Autoencoders, Anomaly Detection, Machine Learning, PyTorch, TensorFlow, Python, Grafana, Retrieval-augmented Generation (RAG), Open-source LLMs, Vector Data, Large Language Models (LLMs)

Machine Learning Engineer

2019 - 2021
Fourthline
  • Oversaw a data annotation strategy, ETL pipelines, and training frameworks for deep learning models to achieve end-to-end identity document authentication and provide calibrated risk scores.
  • Researched and created frameworks for uncertainty estimation, unsupervised learning, and generative modeling to improve fraud detection capabilities.
  • Developed, deployed, and maintained optimized Python APIs for serving ML models.
Technologies: Computer Vision, Python, TensorFlow, Large Language Models (LLMs)

Software Design Engineer

2018 - 2019
ASML Holding
  • Developed model-driven software components and algorithms for new machine modules.
  • Implemented data visualization pipelines to verify new motion algorithms.
  • Developed a code analysis tool to inspect illegal interfaces that reduced development time.
Technologies: Python, System Design, Modeling

Robotics Software Developer

2017 - 2018
Magazino
  • Led the technical onboarding and deployment of autonomous fleets at two intra-logistics warehouses.
  • Developed navigation algorithms based on topological graphs for multi-robot applications.
  • Created a robot-in-the-loop testing pipeline to improve the integration and validation process.
Technologies: Computer Vision

Experience

Account Fraud Detection Using Graph Machine Learning

A graph-based approach to detect fraudulent account holders onboarded on multiple platforms. The graph topology captures how accounts are linked together through transactional patterns or KYC attributes. Viewing this problem from the lens of graphs allows us to capture subtle fraud patterns, significantly improving recall.

Visualizing TensorFlow Metrics in Kibana

https://medium.com/fourthline-tech/how-to-visualize-tensorflow-metrics-in-kibana-761268353ca3
This project explains how to visualize TensorFlow serving metrics in Kibana within the Elastic Stack using Prometheus and Metricbeat. I created a containerized framework that shows how to closely track, monitor, and visualize the performance of a machine-learning model within a deployed service running in a production environment.

Help Navigate Robots: Terrain Classification Using IMU Data

https://github.com/adelizer/kaggle-sandbox/tree/master/terrain-classification
I built an RNN-based classification model that enables robots to recognize the floor surface using data collected from inertial measurement units (IMU sensors). The model can differentiate nine different surfaces with an accuracy of over 70% based on the IMU data.

Unlock Key Insights Using IMF Data Effortlessly!

Use AI and foundational models to stay informed about the latest key economic indicators from around the world. By making use of the International Monetary Fund's data API, you can guarantee "data freshness" that all the data is up to date. The API provides access to thousands of key indicators and time-series data for all countries of the world. This includes but is not limited to the consumer price index, unemployment rates, GDP, and many more.

Education

2015 - 2017

Master's Degree in Computer Science

University of Twente - The Netherlands

2009 - 2014

Bachelor's Degree in Robotics and Mechatronics

German University in Cairo - Cairo, Egypt

Certifications

MARCH 2021 - PRESENT

Build Better Generative Adversarial Networks (GANs)

Coursera

SEPTEMBER 2019 - PRESENT

Deep Learning

Coursera

FEBRUARY 2019 - PRESENT

Machine Learning

Coursera

Skills

Libraries/APIs

TensorFlow, Keras, Flask-RESTful, Pandas, NumPy, PyTorch, Scikit-learn, PySpark

Tools

Git, PyCharm, Kibana, Elastic, Jupyter, Pytest, Grafana

Languages

Python, C++, JavaScript

Platforms

Jupyter Notebook, Docker, Vercel

Frameworks

Flask, Hadoop, Next.js

Paradigms

Design Patterns, Anomaly Detection

Other

Data Analysis, Machine Learning, Data Science, Computer Vision, Software Architecture, Robotics, Retrieval-augmented Generation (RAG), Data Visualization, System Design, Modeling, Prometheus, Time Series Analysis, Robot Operating System (ROS), FastAPI, Artificial Intelligence (AI), Deep Learning, Deployment, Research, Variational Autoencoders, Open-source LLMs, Vector Data, Large Language Models (LLMs)

Collaboration That Works

How to Work with Toptal

Toptal matches you directly with global industry experts from our network in hours—not weeks or months.

1

Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3

Start your risk-free talent trial

Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring