I am a Senior AI Engineer with more than six years of experience in AI research and engineering. I build enterprise AI systems from the first requirement to production. My daily work covers multi-agent systems, RAG and GraphRAG, AI assistants and the automation of ERP and SAP processes with AI. I lead small delivery teams and work directly with clients as their main technical contact. My research background is in machine learning for robotics and in deep graph neural networks. My Master's thesis led to a NeurIPS 2024 paper that I co-authored, and it received the 2025 prize for outstanding scientific achievement from the Freunde der TU Darmstadt. I also wrote an Arabic book that explains AI to readers without a technical background, and I write about AI for the Syrian newspaper Al-Thawra.
About
About me
I build AI systems for companies, I did research on robots and graph neural networks, and I write about AI for people without a technical background.
Profile
Path
How I got here
-
2010
Four semesters of agricultural engineering at Damascus University.
-
2016
Arrived in Germany. Volunteered with the German Red Cross in an emergency shelter for refugees and reached C1 in German within about a year.
-
2018
Started computer science at TU Darmstadt with a full scholarship from the Studienstiftung des deutschen Volkes.
-
2020
Research assistant at the Intelligent Autonomous Systems Lab, working on machine learning for robots.
-
2021
Bachelor's degree with a thesis graded 1.0. Started the Master's in autonomous systems.
-
2022
Joined ponturo consulting in Frankfurt as a Machine Learning Engineer.
-
2024
Master's degree with a thesis graded 1.0. The thesis research became a NeurIPS 2024 paper, which I presented in Vancouver.
-
2025
Senior AI Engineer. The Master's thesis received the prize for outstanding scientific achievement.
-
2026
Published an Arabic book on AI and started writing for Al-Thawra.
Skills
What I work with
- Generative and agentic AI
- LangChain, multi-agent systems, agent workflows and tool use, AI assistants, prompt and evaluation design, model fine-tuning.
- Retrieval and knowledge
- RAG, GraphRAG and knowledge graphs, long context handling, document parsing, summarization, transcription and structured data extraction.
- Machine learning and deep learning
- PyTorch, training and fine-tuning, graph neural networks, computer vision with encoder decoder models, reinforcement learning, classical ML and model evaluation.
- Robotics
- ROS, Gazebo, Isaac Gym, OMPL, robot perception, obstacle avoidance, grasp detection and motion planning.
- Enterprise AI and automation
- AI driven automation of ERP processes, SAP and SAP HANA integration, AI agents for statistical analysis, queries and reporting.
- Software and MLOps
- Python at expert level, FastAPI, Go, Java, JavaScript and Vue.js, cloud native microservices, containers and orchestration, MLOps, CI and full stack web applications.
- Leadership
- Technical team lead, end to end project ownership, requirements analysis, solution architecture, client communication, mentoring and technical hiring.
- Research and writing
- Experimental design, benchmarking, scientific publishing, technical writing and science communication.
Education
TU Darmstadt
-
Oct 2021 to Feb 2024
M.Sc. Autonomous Systems (Autonome Systeme)
TU Darmstadt, Department of Computer Science
My thesis Selective Activations for Addressing Oversmoothing in Deep Graph Neural Networks was graded 1.0 and supervised by Prof. Dr. Kristian Kersting. It received the 2025 prize for outstanding scientific achievement and became the basis of the NeurIPS 2024 paper. The program, taught in German, covers machine learning, robotics and the perception, planning and control of autonomous systems.
-
Oct 2018 to Sep 2021
B.Sc. Computer Science
TU Darmstadt, Department of Computer Science
My thesis An Educational Framework for Robot Learning was graded 1.0 and supervised by Prof. Dr. Jan Peters at the Intelligent Autonomous Systems Lab. I designed and built a Python robot simulator with exercises for students, using Pymunk, Pyglet and Sphinx. The exercises cover robot kinematics with Denavit-Hartenberg, forward and inverse kinematics, splines, PID control, Markov decision processes with policy and value iteration, ProMPs, REPS and ball catching. The program gave me a solid base in computer science, software engineering, algorithms and mathematics.
-
2010 to 2012
Agricultural Engineering, four semesters, not completed
Damascus University
Languages
- Arabic Native
- German C1, TestDaF 19 of 20 points
- English C1