I build intelligent systems that bridge the gap between cutting-edge AI research and real-world impact. From multi-agent architectures to production-grade ML pipelines, I turn complex problems into elegant solutions.
I'm a passionate AI Engineer and Machine Learning Developer currently pursuing my Bachelor's in Computer Science at Ain Shams University. With a strong foundation in building production-ready AI systems, I specialize in creating intelligent multi-agent architectures, knowledge graphs, and scalable ML pipelines.
Currently working at Wider, a multinational company, where I build and deploy backend AI agent infrastructure handling production authentication flows and semantic metadata enrichment using LLM extraction pipelines.
As Head of AI at iCLUB, I lead strategic AI initiatives, organize workshops on generative AI and model deployment, and mentor the next generation of AI developers.
Ranked 2nd in an ML project at Ain Shams University.
Won 1st place for an NLP project at Ain Shams University.
GPA: 3.5 / 4.0 (A-)
Grade: 99.3% | 30 Hours | AI Agents, LLMs, LangChain, RAG, Multi-Agent Systems
Multi-agent LangGraph pipeline (Extractor → Investigator → Resolver → Explainer) for automated claim triage and resolution with semantic deduplication using pgvector HNSW indices and multilingual embeddings.
Terminal-native multi-agent coding pipeline powered by LangGraph. Splits work across specialized agents (Thinker → Worker → Debugger), each with its own prompt, tools, and model provider. Supports Anthropic, Groq, OpenRouter, and NVIDIA — mix providers mid-session.
Automated chatbot integrating multiple flight and hotel reservation APIs for real-time booking with workflow automation pipelines using n8n.
Desktop application for creating and managing exams with automated grading, user authentication, and performance tracking using OOP principles.
Compared ResNet50 (81.9%) vs EfficientNetB0 (96.3%) on NWPU-RESISC45 dataset for multi-class land cover detection using transfer learning.
Multi-class classification using XGBoost achieving 86.6% accuracy with engineered interaction features for improved predictive performance.
Time series forecasting with LightGBM predicting weekly sales (MAE ≈ $7,277) using rolling averages and TimeSeriesSplit cross-validation.
Applied SMOTE balancing and trained Logistic Regression, SVM, and Random Forest achieving 76% accuracy for diabetes prediction.
Segmented customers into 5 clusters via Elbow method with K-Means, visualizing spending vs. income patterns.
Graph-based segmentation with Gaussian smoothing, parallel RGB processing, and multi-format export capabilities.
Multilingual (Arabic/English, multi-dialect) LangGraph sales agent that helps prospective students explore a 52-course catalog, qualifies leads with automatic CRM ticket creation, and notifies the sales team on WhatsApp in real time.
Unified analytics view that consolidated 7 multi-source LMS exports, resolved 37 data-quality issues, and delivered interactive EDA visualizations to surface student engagement and performance trends.
I'm always interested in hearing about new opportunities, collaborations, or just having a great conversation about AI and technology.