I'm an AI Engineer based in Beirut, Lebanon, specializing in production-grade AI systems across computer vision, NLP, speech AI, and robotics. I lead engineering teams and collaborate with CTOs and researchers to design and deploy scalable AI solutions using FastAPI, Docker, gRPC, and Kubernetes.
- MSc. AI & Machine Learning — Lebanese University (2022–2024)
- BSc. Computer Science — Lebanese University (2019–2022)
- Deployed 10+ AI models across vision, NLP, and speech domains
- Led student team to 3rd place in a national AI competition with BaseerAI Smart Glasses
- Robotics Instructor since 2019 at Kalimat Association — mentoring the next generation
- Huawei HCIA-AI v3.0 certified (2022)
- Strong focus on real-time inference, system architecture, and end-to-end AI deployment
| AI / ML | Computer Vision · NLP · Speech AI · OCR · Face Recognition · RAG · Explainable AI |
| AI Deployment | FastAPI · Docker · gRPC · Kubernetes · REST APIs · Model Serving · Real-Time Inference |
| Edge & Robotics | Jetson · Raspberry Pi · Arduino · Hardware-Software Integration · Embedded AI |
| Data Systems | Elasticsearch · Apache Airflow · Neo4j · ArangoDB · MongoDB · MySQL |
| Web & Software | React · Node.js · ASP.NET · TypeScript · JavaScript · MongoDB |
Robotics & Embedded 2019 → Present ──► Arduino · Raspberry Pi · Python (Kalimat Assoc.)
Mobile & .NET 2020 → 2021 ──► Xamarin · ASP.NET Core · Azure (Scinops)
Full-Stack Web 2022 → 2023 ──► React · Node.js · MongoDB (Salem Group)
AI Engineering NOW 2023 → Present ──► CV · NLP · Speech · Docker · K8s · FastAPI
AI Thesis (XAI) 2024 ──► Sobol XAI · Satellite Imagery (GEOAI)
AI for Good 2025 ──► BaseerAI Glasses · 3rd Place (Haret Hreik)
This breadth is my superpower — understanding the full stack from silicon to intelligence makes me a sharper AI engineer and a stronger team lead.
|
AI-powered wearable for visually impaired users. Integrated object detection, OCR, face recognition, and real-time text-to-speech on Raspberry Pi. Led student team to 3rd place in a national AI competition.
|
Multi-modal AI platform spanning speech, text, vision, face recognition, RAG, and deepfake detection. Built with Docker, gRPC, Kubernetes, and Python model-serving APIs for scalable real-time inference.
|
|
Built an Explainable AI framework adapting Sobol sensitivity analysis for pixel-level segmentation of high-resolution satellite imagery. Used entropy filtering and LaMa inpainting validation on the FloodNet dataset.
|
End-to-end pipelines for extracting insights from heterogeneous data sources. Orchestrated with Apache Airflow, indexed via Elasticsearch, and modeled in Neo4j and ArangoDB for graph-based discovery.
|
|
Real-time listener for subscribed Telegram channels — collects, filters, and monitors topics of interest such as AI news, enabling automated tracking and downstream analysis pipelines.
|
Secure file transfer system with authentication, encryption, and audited APIs. Built with WPF on the client side and Python + ASP.NET on the backend for cross-platform office workflows.
|
