Ahmad Droby - Computer Vision & Machine Learning Researcher

Education

2018 - 2023

Ph.D. in Computer Science

Ben-Gurion University of the Negev, Beersheba

Dissertation: Using Weakly Supervised Learning to Solve Visual Computing Problems

Teaching Excellence Award (2022) Council for Higher Education Scholarship (2019-2022) Hi-Tech and Bio-Tech Scholarship (2018)

2016 - 2018

M.Sc. in Computer Science

Ben-Gurion University of the Negev, Beersheba

Thesis: Surface Detection and Deformation Detection in a Video Stream

Council for Higher Education Scholarship (2016-2018) Academic Excellence Award (2018)

2013 - 2016

B.Sc. in Computer Science

Ben-Gurion University of the Negev, Beersheba

2016

Android and Web Application Development

The College of Management, Beersheba

Work Experience

Algorithm Team Lead

Neolithics.ai

2025 - Present

Designed and led development of the Automatic Model Build Flow (Data Collection, Dataset Lifecycle, Model Building, Model Deployment)
Developed advanced computer vision & hyperspectral models for quality control of fresh produce (blueberries, avocados, grapes, garlic)
Built multi-label and ordinal classification models (CNN, ViT, CNN-ViT hybrids) with severity-aware losses
Created reproducibility and deployment standards (TorchScript/ONNX/TensorRT optimizations)
Engineered large-scale training pipelines with PyTorch Lightning, Optuna HPO

Senior Computer Vision Researcher

Neolithics.ai

2024

Led R&D effort turning hyperspectral and RGB imagery of fresh produce into real-time, production-grade quality-control predictions
Designed multimodal fusion network blending RGB and 224-band hyperspectral data
Designed end-to-end ML pipelines: data ingestion, feature extraction, automated validation
Delivered faster, Jetson-ready inference via hybrid CNN/transformer and signal-processing techniques

CTO and Co-Founder

Mirage Dynamics

2022 - 2024

Co-founded Mirage Dynamics and served as CTO, leading integration of AR-style in-video advertising using AI
Developed system for analyzing and replacing existing ads in videos using AI models for object detection, segmentation, and tracking
Built dynamic ad insertion system for embedding ads in live video streams, optimized for high performance in HLS
Created contextual video analysis tools using image/video captioning, NLP, and zero-shot classification
Designed advanced ad insertion system integrating AI with manual processes and Adobe After Effects

Selected Publications

2022

Digital Hebrew Paleography: Script Types and Modes

Ahmad Droby, et al.

Journal of Imaging 8(5): 143

A comprehensive study applying deep learning to classify medieval Hebrew manuscripts into 14 classes based on script style and mode.

2022

Textline Alignment on the Image Domain

Ahmad Droby, et al.

International Journal of Document Analysis and Recognition 25(4): 415-427

Novel approach to text line alignment in document images, improving recognition accuracy.

2022

HST-GAN: Historical Style Transfer GAN for Generating Historical Text Images

Ahmad Droby, et al.

DAS 2022: 523-537

Generative model for creating authentic-looking historical text images using style transfer techniques.

2021

Unsupervised Learning of Text Line Segmentation by Differentiating Coarse Patterns

Ahmad Droby, et al.

ICDAR (2) 2021: 523-537

Innovative unsupervised approach to text line segmentation in handwritten documents.

2021

VML-HD: The Historical Arabic Documents Dataset for Recognition Systems

Ahmad Droby, et al.

Dataset Publication

Comprehensive dataset for advancing Arabic historical document recognition research.

View All Publications on Google Scholar DBLP Profile

Get In Touch

I'm always interested in collaborating on research projects, discussing new ideas, or exploring opportunities in machine learning and computer vision.

Email

Available on LinkedIn

Education

Ben-Gurion University of the Negev

Current Position

Algorithm Team Lead at Neolithics.ai