Mohamad Dbouk | AI Engineer

Transforming Ideas into Intelligent Solutions

AI Engineer specializing in agentic systems, multi-agent architectures, and deep learning. Building intelligent solutions that autonomously reason, plan, and execute complex tasks

Let's get in touch View Projects GitHub

About Me

I'm an AI Engineer specializing in agentic systems and multi-agent architectures. My work focuses on building intelligent agents that can reason, plan, and interact with external tools and APIs.

I have hands-on experience with LangChain, LangGraph, and RAG pipelines, creating solutions that automate complex workflows. I'm comfortable with deep learning frameworks, deploying applications with Docker, and working across the full AI development cycle.

I'm passionate about solving real problems with AI and continuously learning new technologies in this fast-evolving field.

Years Experience

AI Projects

15+

Technologies

100%

Passion

Technology Stack

AI & Machine Learning

Python (Advanced)
TensorFlow / PyTorch
Scikit-learn
OpenCV
Hugging Face Transformers
Custom Neural Networks
Langchain
Langgraph

Agentic Systems

LangChain (Expert)
LangGraph
AutoGen
Agent Frameworks
Multi-Agent Systems
RAG Implementation

Development Tools

Git & GitHub
Docker
FastAPI / Flask
Jupyter Notebooks
MLflow
Cloud Platforms (AWS)

Featured GitHub Projects

Deep Learning Anomaly Detection

A deep learning–based anomaly detection project focused on identifying unusual patterns in complex datasets using neural networks. Designed to detect rare events and outliers through representation learning and reconstruction error analysis.

Python TensorFlow / Keras NumPy Pandas Scikit-learn Matplotlib

Key Features

Deep learning models for anomaly detection using reconstruction-based techniques
Feature engineering and data preprocessing for high-dimensional data
Threshold-based anomaly scoring and detection
Model evaluation using reconstruction error analysis
Visualization of anomalies and performance metrics
End-to-end ML workflow from data preparation to evaluation

code

RAG Chatbot (Claude API + LangChain + Gradio)

A Retrieval-Augmented Generation (RAG) chatbot that enables interactive question-answering over PDF documents using a local document ingestion, embedding, and FAISS-based vector retrieval pipeline, with a configurable LLM backend supporting Claude API and external providers. The system combines vector search with LLM reasoning and provides both a modern web interface and CLI support.

Python LangChain FAISS Claude API Gradio Hugging Face Transformers Docker / Docker Compose

Key Features

End-to-end RAG pipeline (PDF ingestion → embedding → retrieval → generation)
Local document processing and vector storage for private, efficient retrieval
Pluggable LLM backend supporting Claude API and other external providers
Vector-based semantic search with FAISS and persistent on-disk indexing
Interactive Gradio web UI for chatting with documents and managing the knowledge base
Command-line interface (CLI) for ingestion, querying, and automation workflows
Dockerized deployment for reproducible, platform-independent execution
Smart PDF deduplication using metadata to prevent redundant processing
Modular and configurable architecture designed for extensibility and scalability

code demo

Credit Card Fraud Detection

A machine learning project focused on detecting fraudulent credit card transactions using anomaly detection and classification techniques to handle highly imbalanced financial datasets.

Python Scikit-learn Pandas NumPy Matplotlib Seaborn

Key Features

Fraud detection on highly imbalanced transaction data
Feature scaling and preprocessing for financial datasets
Anomaly detection and supervised ML models
Evaluation using precision, recall, F1-score, and ROC-AUC
Visualization of fraud patterns and model performance
Real-world fraud detection workflow and risk analysis

code

Titanic ML Project

A classic supervised machine learning project predicting passenger survival on the Titanic using structured data analysis and multiple classification models.

Python Scikit-learn Pandas NumPy Matplotlib Seaborn