resume
This page will contain my career resume. Last updated 23/07/2026
Contact
| Name | Ahmed Atef Ahmed Aly |
| Label | Research Engineer |
| ahmed.aly@mbzuai.ac.ae | |
| Personal | ahmedatefahmedaly@gmail.com |
Work
-
2026.06 - Present Abu Dhabi, UAE
Research Engineer I
Mohamed bin Zayed University of Artificial Intelligence, BioMedIA Lab
Echocardiography foundation models research under Prof. Mohammad Yaqub
- Researching echocardiography foundation models, extending view-aware pretraining toward temporal-causal objectives and automated clinical report generation.
- Supervising a team of four interns on a multimodal vision-language model for medical image parsing (FLARE 2026, MICCAI).
-
2025.05 - Present Abu Dhabi, UAE
Machine Learning Engineer (Part-time)
Labib AI
AI solutions for enterprise proposals and on-device deployment
- Designed AI solutions for government and enterprise proposals: visitor analytics, predictive maintenance, and multi-entity inspection ecosystems.
- Benchmarked multi-model edge AI workloads (multi-stream YOLOv8, LLMs, VLMs, Whisper) on Qualcomm Dragonwing IQ-9075 for on-device deployment.
- Engineered a real-time surgical room understanding pipeline integrating Whisper-3, Qwen-2.5, and HRNet, demoed to healthcare professionals via a Django web application.
-
2023.01 - Present Abu Dhabi, UAE
Computer Vision Engineer, Co-founder
Koralyze AI
Football analytics from broadcast footage
- Co-founded a football analytics company generating tracking and event data directly from broadcast footage for downstream analytics.
- Presented at GITEX Expand North Star 2023; leading partnership development with UAE sports federations.
- Won first place in the UAE IoT & AI Challenge startup category (2023); part of the Khalifa Innovation Center incubation program.
-
2020.08 - 2024.05 Sharjah, UAE
Education
-
2024.08 - 2026.05 Abu Dhabi, UAE
MSc
Mohamed Bin Zayed University of Artificial Intelligence
Computer Vision
- Thesis: EchoVision, a vision-language foundation model for echocardiography using view-aware contrastive pretraining to address view-text misalignment.
- Trained on 4M echocardiography videos (36 TB) using distributed multi-GPU SLURM pipelines; outperforms EchoCLIP on all nine internal regression tasks and PanEcho on six.
-
2016.08 - 2020.05 Abu Dhabi, UAE
BSc
Khalifa University
Applied Mathematics and Statistics
- Optimization of global minima in nonlinear physical systems using metaheuristic statistical models.
Skills
| Medical AI Foundation Models | |
| Self-Supervised Learning | |
| CLIP/JEPA Contrastive Learning | |
| Knowledge Distillation | |
| Vision-Language Models | |
| Medical Imaging (Ultrasound, MRI) |
| Programming & Tools | |
| Python | |
| SQL | |
| C++ | |
| PyTorch | |
| PyTorch Lightning | |
| OpenCV | |
| Docker | |
| Git | |
| Linux | |
| SLURM |
| Languages | |
| English (IELTS 8.0) | |
| Arabic (Native) |
Projects
- 2025.06 - 2025.10
CardioBench
Do Echocardiography Foundation Models Generalize Beyond the Lab? (IJCAI-ECAI 2026)
- Co-authored a standardized benchmark evaluating echo foundation models across nine tasks and eight public datasets.
- 2025.05 - 2025.06
SynSpineMS
Multiple Sclerosis Spinal Cord Lesion Detection from MultiSequence MRIs (MICCAI challenge)
- Led our team's participation in the MICCAI MS-Multi-Spine challenge; finished 2nd overall.
- 2024.10 - 2025.01
Language and Planning in Robotic Navigation
Vision-Language Navigation for multi-step planning in indoor robotic environments (LM4Plan Workshop, AAAI 2025)
- Evaluated across multilingual (Arabic/English) instruction settings.
- 2025.01 - 2025.04
NeuroGNN
Diagnosing Autism Spectrum Disorder using Multimodal Brain Connectivity Network
- Developed a graph-based fMRI analysis pipeline for autism classification using Graph Attention Networks (GATs), modeling brain functional connectivity across seven subnetworks.
- 2025.01 - 2025.04
Perception Challenge for Bin-Picking
Robust pose estimation for the most challenging industrial parts in the 2025 BOP Challenge
- Developed a computer vision pipeline for bin-picking automation, integrating object detection, pose estimation, and grasping strategies.