An AI R&D Engineer based In France
A passionate AI Research and Development Engineer based in France
I’m an AI R&D Engineer with a Ph.D. specialized in AI for Business Process Management (BPM). I combine deep research expertise with hands-on engineering experience, focusing on AI-driven solutions, natural language processing, and generative AI to create scalable, real-world impact.
• Built an AI pipeline with LangGraph and LangChain to auto-generate executable business processes, with LangSmith for monitoring and Azure Foundry for LLM hosting.
• Developed a Spring Boot engine on Flowable for process execution and orchestration.
• Created a workflow API server enabling external integration and dynamic workflow deployment.
• Dockerized all services and automated Azure deployment using GitHub Actions CI/CD.
Skills: LangGraph, LangChain, LangSmith, FastAPI, Azure Foundry, Spring Boot, Flowable (Java), Docker, GitHub Actions, LLM, Prompt Engineering.
• Developed an extension to BPMN for simultaneous modeling of control and data flows, enhancing process efficiency and data integration.
• Built a scalable, web-based process modeling tool using Eclipse GLSP, containerized with Docker, and deployed on Azure for high availability.
• Developed an AI-powered tool that transforms unstructured text into process models using LLMs, and deployed it on AWS for scalable use.
• Led 3 AI-driven projects, mentoring and managing a team of 8 students, focusing on the application of LLMs for BPM and data management automation.
• Presented AI innovations and BPM applications at international conferences and webinars to global audiences.
• Contributed to the R&D tax credit (CIR) process by preparing comprehensive research documentation, ensuring compliance with regulatory standards.
Skills: Java (Eclipse GLSP), Bonita, BPMN, python (Flask, Langchain), Large Language Models (LLMs), nodejs (Vite Framework), Docker, Azure, AWS, JFrog, Git/GitHub.
• Designed and developed fintech solutions, including a CRM system integrated with MetaTrader5 API and a secure E-Wallet platform, enhancing trading operations and digital payments.
• Built synchronized Admin Portals and Mobile Apps, ensuring seamless interaction between brokers, traders, and back-office staff.
• Implemented advanced KYC, OTP verification, biometric login, encrypted transactions, audit trails, and compliance features aligned with financial regulations.
• Deployed applications on secured cloud infrastructure with HTTPS enforcement, SSL/TLS encryption, domain management, and real-time notifications via Firebase Cloud Messaging.
Skills: MetaTrader5 API, Mobile App Development, KYC/OTP, Biometric Login, Firebase Cloud Messaging, Laravel Reverb, Pusher, Cloud Infrastructure (Hostinger VPS, IONOS), SSL/TLS, Encryption, Flutter, Laravel.
• Automated a DataOps pipeline for face extraction from videos, utilizing advanced face detection, cropping, and super-resolution techniques.
• Fine-tuned generative models (StyleGAN2) on curated datasets, achieving high-fidelity cartoon face generation with GCP-based GPU acceleration.
• Developed and deployed a Flask-based web app for dynamic face generation, optimizing model performance on Google Cloud Platform.
Skills: Python (Pytorch, Flask), Google Cloud Platform (GCP), StyleGAN, SemanticStyleGAN, Super Resolution, DataOps, Google Colab, Vertex AI.
• Administered server operations and network systems, ensuring reliable IT infrastructure to support hospital-wide operations.
• Provided technical support and troubleshooting, maintaining seamless IT functionality to support critical healthcare services.
• Developed applications for lab data integration and database conversion, automating workflows and reducing manual effort by 100%.
Skills: C#, SQL, VMware ESXi, Windows Server, MS Access, Networking.
• Created scalable process discovery solutions using Neo4j and Cypher, improving real-time business process analysis.
• Designed graph-based models for object-centric event data, enhancing data integration and performance over traditional systems.
Skills: Python (PM4PY), Neo4j (Cypher).
- Thesis title: “A Comprehensive Framework for Low-Code Data-Centric Process Management”
- Thesis CIFRE collaboration: Telecom SudParis and Bonitasoft
- Thesis Supervisors: Jonathan Baudot, Dr. Benjamin Dalmas and Prof. Walid Gaaloul
- Thesis title: “Enabling Process Mining on Graph Database”
- Thesis collaboration: Lebanese University and Telecom SudParis
- Thesis supervisors: Dr. Nous Assy and Prof. Walid Gaaloul
Machine Learning (K-NN, K-mean, PCA, Decision Tree, Naïve Bayes, HMM, and others), Deep Learning (CNN, RNN, LSTM, GANs, landmarks, style transfer, transfer learning), LLMs (OpenAI GPT-3.5/4, BERT, T5, Ollama, RAG), LangChain, Graph Theory, Process Mining, MLflow, PyTorch, Tensorflow, RAG, Agent AI, LangGraph, numpy, scikit-learn, NLTK, spaCy, openCV.
Hadoop Ecosystem, Spark, Kafka.
Java, Python (Flask, FastAPI), C#, C/C++,Vite.js, Cypher, php (laravel), dart (flutter).
Docker, Git, Azure, AWS, JFrog, GCP, CI/CD.
SQL, MySQL, NoSQL (Firebase and Neo4j), MS Access
I built a Telegram bot that provides real-time weather updates using LangGraph and the Model Context Protocol (MCP). The bot interprets user queries, fetches data from a weather API, and responds naturally in chat.
As part of the project, I tested two agent types: a ReAct-style agent, which reasons step by step while calling tools, and a stateful graph agent, which manages multi-turn conversations with context memory. This comparison highlights different approaches to designing conversational AI flows.
Technologies: Python, LangGraph, MCP, Telegram Bot API, Weather API, LLMs.
GitHub: https://github.com/NourEldin-Ali/mcp-langgraph-weather-telegram
For this project, I focused on building a context-aware Retrieval-Augmented Generation (RAG) chatbot using modern AI frameworks and tools. Leveraging Large Language Models (LLMs), vector databases, and retrieval pipelines, the chatbot delivers accurate and contextually relevant answers from unstructured data sources. The system is architected using Python, LangChain, FastAPI, and Streamlit, combining LLM-driven natural language processing with real-time user interaction.
I used LangChain to orchestrate the flow between document retrieval and LLM prompting. FastAPI serves as the backend API layer, managing interactions between the user interface, the vector database (ChromaDB), and the LLMs. By embedding documents into vector space using LangChain’s retrieval pipeline and storing them in ChromaDB, the system performs fast and efficient similarity searches based on user queries. This architecture ensures responses are grounded in retrieved content, increasing both accuracy and reliability.
On the frontend, Streamlit was used to develop a clean, user-friendly interface for real-time question answering. Users can input natural language queries, which are then processed through the backend pipeline. Retrieved documents are passed to the LLM, which synthesizes a response based on the content. The system is compatible with both OpenAI models and local LLMs via Ollama, supporting models like Mistral and LLaMA for flexible deployment.
This project demonstrates my ability to integrate retrieval mechanisms with generative models, handle vector database operations, and build end-to-end applications that combine multiple AI components. It highlights my skills in prompt engineering, API development, and full-stack LLM applications. The modular and scalable design allows easy adaptation for enterprise-level knowledge assistants or internal document Q&A tools.
Technologies: Python, FastAPI, Streamlit, LangChain, ChromaDB, Ollama, OpenAI, Large Language Models (LLMs).
For this project, I focused on leveraging cutting-edge technologies like Large Language Models (LLMs) and implementing prompt engineering techniques to automate business process modeling. Using Python, Flask, and Langchain, I built a hybrid system that integrates LLM-based text analysis with a structured algorithmic approach. This approach decomposes the complex task of generating business process models from textual descriptions into manageable sub-tasks. The LLM analyzes, clarifies, and extracts process entities such as activities and dependencies, while the modular pipeline provides users with an interactive platform for validating or updating extracted information, resulting in highly accurate BPMN models.
Additionally, I utilized the power of Python and Langchain to integrate the LLM's natural language understanding capabilities within a scalable web framework developed using Flask. This hybrid system not only improves efficiency by reducing token usage but also enhances the accuracy and relevance of generated models. By breaking down the process modeling into distinct steps, this project showcases my skills in prompt engineering, Python development, and the use of AI technologies to solve real-world problems, offering an innovative approach to handling incomplete or non-expert descriptions in process automation.
Furthermore, the project extends its capabilities to support the transformation of textual descriptions into Unified Modeling Language (UML) diagrams. The system extracts structural and behavioral elements from the text, such as classes, attributes, methods, and their relationships, to automatically generate accurate UML diagrams, including class diagrams and sequence diagrams. This functionality aids in visualizing complex software architectures and enhances the clarity of system designs derived from natural language inputs.
In addition to BPMN and UML generation, the system is designed to transform textual data into business data models that can be directly used to execute BPMN workflows. By mapping the extracted entities and process details into executable business data formats, the project facilitates seamless integration with BPMN engines, thereby enabling automated process execution. This capability ensures that the generated models are not only descriptive but also operational, bridging the gap between process design and execution.
Technologies: Python, Flask, Langchain, Large Language Models (LLMs), BPMN, nodejs (Vite Framework), AWS, plantUML.
This E-Wallet system is a comprehensive digital payment solution comprising an Admin Portal, a Mobile App, and a Backend developed using Laravel. It supports money transfers, currency exchanges, product purchases, and secure user authentication via KYC and OTP mechanisms.
Admins can manage users, transactions, currencies, trading groups, and system settings, ensuring compliance with regulations. Key features include transaction management, fee handling, KYC approval, and integration with payment gateways and Firebase Cloud Messaging for real-time notifications.
Clients and agents can manage their accounts, perform internal transfers, request withdrawals, deposit via credit card, and create real or demo accounts. Biometric authentication, QR code payments, and OTP provide enhanced security.
The Laravel backend manages user data, KYC verification, OTP generation, and secure transaction processing. It integrates real-time currency exchange, product and merchant management, and ensures regulatory compliance through encryption and robust security protocols.
This E-Wallet system is a robust solution designed to handle digital payments, currency exchanges, product purchases, and secure onboarding through KYC and OTP, ensuring high levels of security and regulatory compliance for users, agents, and administrators.
Technologies: Flutter, Laravel, Firebase Cloud Messaging, MySQL.
This CRM system is designed for managing trading accounts and facilitating financial deposits and withdrawals, directly integrated with MetaTrader5 via API. It provides tools for both clients and brokers to manage their accounts and transactions securely and efficiently. The backend is powered by Laravel for robust data management and security.
Technologies: Flutter, Laravel, MetaTrader5 API, MySQL.
I developed this website for Innovatech. It is a clean, responsive static site designed to showcase the company's services and brand identity. The website incorporates modern web technologies and ensures cross-platform compatibility, providing an optimal user experience on both desktop and mobile devices.
Technologies: HTML5, CSS, JavaScript, PHP (for email functionality).
This project is an open-source framework designed to facilitate the modeling and manipulation of DF-BPMN compliant workflows which is an extension of BPMN 2.0. Built on the Eclipse Graphical Language Server Platform (GLSP), it offers a flexible Java-based metamodel for generating, importing, and exporting BPMN models, as well as supporting the extension of BPMN elements. This project includes both a client-side graphical modeling tool and a server-side component that allows for seamless integration with modern web technologies. It's an ideal tool for developers looking to build or extend BPMN 2.0 workflows within custom applications, making it highly adaptable for business process automation scenarios
Technologies: Java (Eclipse GLSP), BPMN, Typescript
Source Code: https://github.com/NourEldin-Ali/open-bpmn
I fine-tuned StyleGAN to generate high-quality cartoon faces. Using a custom dataset, I adapted the model to produce diverse cartoon images with distinct features. Moreover, I modified the original StyleGAN architecture to incorporate conditional inputs, allowing precise control over the generated faces. This enhancement enables the generation of cartoon faces with specific attributes, such as gender (male or female), age (young or old), and skin tone. These customizations were achieved by manipulating the latent space and refining the model to respond to these conditions, ensuring the output meets the desired criteria.
Indeed, I modified facial emotions in generated images by manipulating the latent space directions of StyleGAN. I explored specific vectors in the latent space that correspond to various emotions like happiness, sadness, anger, and surprise. By adjusting the direction of these vectors, I was able to seamlessly change the emotional expression of generated faces while maintaining their core identity and other attributes, such as gender and age. This approach allows for precise and dynamic control over facial emotions in generated images, demonstrating the power of StyleGAN's latent space manipulation.
Technologies: Python, Pytorch, StyleGAN2, Google Colab, Google Cloud Platform.
I developed a workflow to automate the process of extracting and enhancing cartoon faces from video files. The workflow begins by downloading cartoon movies, followed by detecting all faces in the video frames. Using advanced face detection algorithms, the process specifically targets cartoon faces, ensuring they are aligned to a specific position. Each face is then cropped to a predefined resolution. Additionally, I integrated a super-resolution model to upscale and enhance the quality of the extracted faces, ensuring high-resolution output. This end-to-end pipeline efficiently processes cartoon movies, resulting in high-quality, precisely cropped facial images suitable for further applications.
Technologies: Python, Google Cloud Platform, Super Resolution Model.
I employed Semantic StyleGAN to perform targeted facial transformations by modifying specific facial segments. Leveraging semantic manipulation in the latent space, I was able to alter individual features such as hair style, hair color, and add facial hair (e.g., beards or mustaches) without affecting other aspects of the face. By isolating and adjusting these distinct facial regions, the transformations maintained a natural and cohesive appearance. This approach enables high flexibility in customizing specific facial attributes while preserving the overall identity of the generated face.
Technologies: Python, Pytorch, Semantic StyleGAN, Google Colab, Google Cloud Platform.
I developed a fully functional handmade incubator using an Arduino microcontroller. The primary goal of this project was to create a system that maintains optimal conditions for egg incubation, specifically controlling temperature and humidity, while incorporating an automatic egg-turning mechanism. This incubator offers a low-cost, DIY solution that mimics the environmental conditions of a traditional incubator, making it ideal for hobbyists and small-scale poultry breeders.
Components: Arduino Uno, DHT22 Senso, Servo Motor, LCD Display, Relay Module, Power Supply.
The LifeSaver mobile application is a cross-platform tool designed to simplify the process of blood donation, making it more accessible and efficient. Built using Flutter, an open-source framework developed by Google, the app allows users to register as blood donors or request blood donations through a seamless interface available on both Android and iOS. Flutter, using the Dart programming language, was selected for its efficiency in creating high-performance apps with a single codebase for both platforms. The integration of Flutter ensures that the application delivers a smooth and native-like experience while maintaining consistency across devices.
The LifeSaver app addresses the challenge of timely communication between hospitals and potential blood donors. It connects users to hospitals in need of blood donations and allows them to respond to requests, track their donation history, and earn points based on their contributions. The app also features real-time notifications, allowing hospitals to post urgent requests and donors to be alerted instantly based on their proximity to the healthcare centers. This system reduces the delays often experienced in traditional blood donation processes, making the app a critical tool in saving lives
Technologies: Flutter, Dart, Firebase Cloud Messanging.
Source Code: https://github.com/NourEldin-Ali/Blood-Donation
I developed an Android application titled "School Student Management." The project aimed to enhance the interaction between students, teachers, and administrators by streamlining class management and communication. The application allows teachers to assign tasks, students to view and complete their assignments, and administrators to manage user accounts, classes, and courses. The technical backbone of the project involved using Java for coding, XML for design, and Firebase for database management, messaging, and notifications. This solution simplifies the educational process by creating an efficient and user-friendly interface for both students and teachers, while also providing administrative control over key educational functions.
Technologies: Android Java, Firebase Cloud Messanging, Firebase.
Date: February 2025
I have successfully completed all four courses in the Natural Language Processing Specialization by DeepLearning.AI on Coursera.
Completed Courses:
Certificate Link: https://coursera.org/verify/professional-cert/YKL2MJTK0BEK
Date: February 2025
I have successfully completed all four courses in the MLOps | Machine Learning Operations Specialization by Duke University on Coursera.
Completed Courses:
Certificate Link: https://coursera.org/verify/professional-cert/SHSDCYDORDWF
Date: January 2025
I have successfully completed all three courses in the PyTorch Ultimate 2024 - From Basics to Cutting-Edge Specialization by Packt on Coursera.
Completed Courses:
Certificate Link: https://coursera.org/verify/professional-cert/E2EMBY87JKTC
Date: Decembre 2024
I have successfully completed all three courses in the Machine Learning Specialization by Andrew Ng (Coursera).
Completed Courses:
Certificate Link: https://coursera.org/verify/professional-cert/E2EMBY87JKTC
Date: Novembre 2024
I have successfully completed all five courses in the Deep Learning Specialization by Andrew Ng (Coursera).
Completed Courses:
Certificate Link: https://coursera.org/verify/professional-cert/2IFHVU55O343
Date: January 2024
I completed a three-course specialization in Prompt Engineering offered by École Polytechnique through Coursera. The courses focused on mastering techniques for crafting prompts that leverage generative AI tools like ChatGPT. I learned key techniques including advanced prompt patterns, automating tasks like data analysis, generating personalized content, and solving problems with AI-driven solutions. The hands-on projects enhanced my ability to apply these models in real-world contexts, from automating processes to generating creative solutions. This experience has significantly boosted my AI literacy and productivity skills.
Courses:
Certificate Link: https://coursera.org/verify/specialization/ELSPHXD3BWPF
Date: January 2024
I have successfully completed 3 out of 6 courses in the Google Project Management Certificate on Coursera. These courses provided me with a strong understanding of the key phases in project management and equipped me with practical skills to effectively lead and manage projects.
Completed Courses:
While I have not yet completed the remaining three courses, the knowledge and skills acquired from the completed courses have provided a strong foundation in project management principles.
Date: March 2022
I have successfully completed the Google IT Support Professional Certificate, a program consisting of five comprehensive courses designed to build expertise in IT support. Through this program, I developed hands-on skills in troubleshooting, system administration, networking, operating systems, security, and customer support, preparing me for roles in technical support and IT operations.
Completed Courses:
Certificate Link: https://coursera.org/verify/professional-cert/W2QKPEQMYBSR
Date: August 2021
I have completed a comprehensive set of six courses through the Neo4j Academy, gaining hands-on experience with graph databases and Cypher query language. These courses include topics such as querying with Cypher, creating nodes and relationships, using indexes for query optimization, and best practices in Neo4j 4.x. Additionally, I gained skills in importing data into Neo4j and designing graph data models tailored to specific use cases. My training provides me with a solid foundation to design, implement, and optimize graph-based applications efficiently using Neo4j.
Certificates Number:
Date: July 2019
I have successfully completed the Cisco CCNA1 Routing and Switching: Introduction to Networks course, conducted onsite at the Lebanese University. This introductory program provided me with a solid foundation in basic networking concepts, including network architecture, protocols, and services. I gained hands-on experience in configuring routers and switches, as well as troubleshooting basic network issues. This course introduced me to essential LAN and WAN technologies, preparing me for further studies in networking and modern network infrastructures.
I specialize in delivering innovative AI solutions that help businesses stay ahead of the curve.
With years of experience in AI technologies, I specialize in delivering innovative AI solutions that help businesses stay ahead of the curve. My hands-on experience with cutting-edge technologies like Generative Adversarial Networks (GANs), StyleGANs, and Large Language Models (LLMs) enables me to provide advanced AI consulting and development services tailored to your needs. I’ve worked on projects where LLMs were employed to convert text into Business Process Model and Notation (BPMN), transforming complex business processes into efficient, automated workflows. My deep technical knowledge allows me to create AI-powered solutions that automate tasks, improve decision-making, and deliver tangible business outcomes.
I Having worked across various industries, I understand the intricacies of applying AI in real-world contexts. Whether it’s automating repetitive tasks or enhancing data-driven decision-making, my solutions are designed to seamlessly integrate into your operations. Every AI project I undertake is rooted in my passion for innovation and my commitment to delivering results that drive real value for your business.
Key AI Projects:
I bring a wealth of experience in software development, having led and executed various projects across multiple platforms.
I bring a wealth of experience in software development, having led and executed various projects across multiple platforms. My expertise spans mobile apps, desktop applications, and enterprise software, giving me a comprehensive understanding of what it takes to create functional, scalable, and user-friendly solutions. From developing a highly secure E-wallet app to building a comprehensive CRM system, my experience enables me to tackle complex challenges with ease. I’ve also developed a blood donation mobile app, a school management system, and a client archiving solution, showcasing my versatility and ability to deliver tailored software solutions across different sectors.
With every project, I focus on understanding the unique needs of the business to create custom software that drives efficiency, enhances user engagement, and supports business growth. My hands-on approach ensures that each solution is meticulously designed, rigorously tested, and deployed with precision. When you work with me, you’re not just getting a developer—you’re partnering with someone who is dedicated to delivering a product that exceeds your expectations and helps your business succeed.
Key Software Development Projects:
I offer expert IT consultation services that optimize and enhance your systems
With a rich background in IT infrastructure and cloud computing, I offer a full range of IT Consultation services designed to optimize your technology stack and ensure seamless operations. My expertise spans traditional networking environments like ESXi and Windows Server, cloud platforms such as AWS and Google Cloud Platform (GCP), and hosting solutions for web applications. Whether you need help managing on-premise systems, migrating to the cloud, or optimizing your cloud infrastructure, I provide tailored solutions that ensure your IT systems are secure, scalable, and efficient.
I have successfully integrated complex IT systems, enabling seamless communication between different technologies and platforms. For example, I have integrated hospital lab machines with hospital management systems, automating data entry and enhancing accuracy. Additionally, I specialize in troubleshooting and resolving IT issues to minimize downtime and ensure smooth business operations. Whether it’s setting up secure hosting environments, optimizing network infrastructure, or connecting systems to ensure smooth data flow, I work to ensure that your IT systems are not only optimized for today but also prepared for future growth.
Key IT Projects:
