Hello World, I'm
Qays Abu Mahfouz.
Freelance AI Engineer & AI Manager
Building intelligent Computer Vision, NLP, LLM integration, Automation, and Real-Time Analytics solutions that drive tangible results.
About Me
I’m Qays , an AI Engineer and AI Manager based in Amman, Jordan. Over the past 3 years, I’ve designed and deployed vision-based security systems, OCR pipelines handling 1.5M+ invoices, and predictive models for healthcare imaging. I excel at turning complex AI research into robust production solutions.
My passion lies at the intersection of Computer Vision, NLP, and real-time analytics. I’ve collaborated with global teams to build training programs, optimize recruitment with AI-driven screening, and automate workflows—saving hundreds of hours per month.
I believe the most exciting AI applications lie at the intersection of multiple domains. Currently, I'm exploring quantum-inspired neural networks and developing models that can reason about ethical dilemmas in automated decision-making systems.
Here are some technologies and frameworks I've been working with recently:
Professional Journey
- Built & rolled out AI training programs adopted by global teams, training hundreds of leaders, executives, and managers.
- Optimized recruitment with AI-driven screening & evaluation, reducing time-to-hire by 30%.
- Automated 20+ workflows via Zapier/N8N, saving 200+ hours/month.
- Consistently exceeded CSAT targets through expert problem-solving & communication.
- Introduced data-driven insights into CRM usage, improving first-call resolution by 15%.
- Mentored 50+ students in data science & deep learning fundamentals.
- Guided 10+ capstone projects 90% success rate.
- Built diamond-price prediction & music-genre classification models (85%+ accuracy).
- Mastered Pandas, NumPy, Matplotlib, Seaborn, CNNs, DNNs & NLP pipelines.
AI Systems
Building Tomorrow's Intelligence
Eagle Eye AI
Real-time retail surveillance & theft prevention using multi-camera vision. with 95% detection accuracy.
License Plate Recognition
Streamlit interface with YOLO & TrOCR for real-time LPR, tracking entries/exits at 98% accuracy.
OCR Invoice Processing
Pipeline handling 1.5M+ invoices via Streamlit + TrOCR, storing parsed data in SQL for real-time dashboards.
COVID-19 Predictive Analytics
Transfer-learning (VGG-19, DenseNet121) on X-rays/CTs, enabling 20% faster pattern detection.
AI Evolution Timeline
Medical Imaging AI
Developed deep learning models for early detection of neurological conditions from MRI scans, achieving 92% sensitivity while reducing false positives by 60%.
Financial Forecasting Network
Created a hybrid time-series model combining transformer architectures with attention mechanisms to predict market fluctuations, outperforming industry benchmarks by 18%.
Multimodal Recommendation Engine
Built an AI system integrating user behavior, content analysis, and contextual information to deliver highly personalized recommendations with a 43% increase in user engagement.
Quantum-Inspired Neural Nets
Pioneered neural network architectures inspired by quantum computing principles, enabling efficient processing of complex probabilistic relationships in data.
Featured Projects
NeuroVision
An AI system that can detect and classify neurological conditions from standard MRI scans with accuracy comparable to specialist radiologists. The model incorporates attention mechanisms to highlight areas of concern and explain its reasoning.
EthicalAI Framework
An open-source framework for evaluating and mitigating bias in machine learning models. Includes tools for fairness assessment, explanation generation, and automated refinement of models to reduce unintended discrimination.
Other Noteworthy Projects
Sentiment Analysis Engine
A fine-tuned transformer model that analyzes sentiment in customer feedback with 94% accuracy across multiple languages and domains, including detecting sarcasm and mixed emotions.
Autonomous Drone Vision
Real-time computer vision system for autonomous drones that enables navigation in GPS-denied environments using solely visual cues and reinforcement learning policies.
ML Operations Platform
A comprehensive MLOps platform for automating the deployment, monitoring, and maintenance of machine learning models in production environments with automatic drift detection.
Time Series Forecasting
Advanced forecasting models combining traditional statistical methods with deep learning approaches for multi-horizon predictions in energy consumption, achieving 28% lower error rates.
Graph Neural Networks for Drug Discovery
Applied GNNs to model molecular structures and predict drug-protein interactions, accelerating the early stages of pharmaceutical research and drug repurposing efforts.
Knowledge Distillation Framework
Created techniques to distill knowledge from large, computationally expensive models into smaller, more efficient ones while maintaining 96% of the original performance.
Get In Touch
I'm currently open to new opportunities and collaborations in AI research and development. Whether you have a project in mind, a research question, or just want to connect, my inbox is always open.
qaysq28@gmail.com