Amina Bouaziz
AI/ML Engineer and Applied Data Scientist with Engineering and Master’s degrees in Computer Science, currently pursuing AI opportunities in Riyadh. Specializes in production-ready LLM applications using LangChain, CrewAI, RAG, and OpenAI APIs. Delivered 10+ AI solutions for clients across HR,retail,etc, including multi-agent systems, NL-to-SQL, and BI automation. Strong research background with IEEE and Springer 2025 publications, plus hands-on expertise in Docker, Kubernetes, FastAPI, and n8n.
Amina Bouaziz
Summary
AI/ML professional focused on transforming real-world business challenges into scalable AI solutions. Experienced in designing and deploying production-ready LLM applications, RAG pipelines, multi-agent systems, NLP solutions, and AI automation, with an emphasis on business impact, system reliability, and practical deployment.
What I Do Best
Key areas of practice spanning data science, AI/ML, research, and software engineering.
AI & Software Engineering Research
Doctoral research at Laval University on leveraging LLMs for automated decomposition of software applications into microservices. Authored and presented an IEEE SANER 2025 paper analyzing microservices deployment trends on Stack Overflow using topic modeling.
NLP & Topic Modeling
Applied NLP techniques including BERTopic, TF-IDF, embeddings, and Hugging Face Transformers for topic modeling, sentiment analysis, and text classification. Experienced in both manual and LLM-assisted evaluation methodologies.
Generative AI & LLM Application Development
Building LLM-powered applications using LangChain, CrewAI, RAG, and OpenAI APIs. Developed multi-agent systems, autonomous agents with memory, and evaluated models through manual and LLM-assisted methods.
Scientific Image Processing
Research experience in X-ray image processing at Sherbrooke University using the novel SSRT method and astra toolbox for non-destructive imaging and 3D object visualization.
Reproducible Research & Tooling
Developed replication packages for empirical software engineering studies using Python, BERTopic, and the Stack Exchange Data Explorer. Committed to reproducible research practices and open science.
Machine Learning & Data Science
Building ML models with Python, scikit-learn, TensorFlow, and Keras for classification, crop detection, and stress detection. Experienced with both traditional ML and deep learning approaches.
Technical Skills
A toolkit refined across data science, programming, web, and design disciplines. Filter by category to focus.
AI & ML
Data Science & Analytics
Database
Programming Languages
MLOps & DevOps
Operating Systems
Web Development
Software Testing
Design
Scientific Computing & Imaging
Certifications
Industry-recognized certifications across AI, deep learning, web development, and design.
Generative AI with Large Language Models
Full-Stack Web Development with React Specialization
Motion Design
UI/UX Design Specialization
Work Experience
Click any role to expand the details. A condensed timeline of where I've made an impact.
- Conducted doctoral research on leveraging Large Language Models (LLMs) for the automated decomposition of software applications into microservices, developing novel approaches to enhance scalability and modularity in software architecture.
- Authored and presented the paper 'An Empirical Study on Microservices Deployment Trends, Topics, and Challenges in Stack Overflow' at IEEE SANER 2025, in collaboration with Professors Mohamed Aymen saied, Mohammed Sayagh, Ali Ouni and Mohamed Wiem Mkaouer.
- Applied topic modeling with BERTopic to analyze Stack Overflow discussions, revealing practical microservices adoption insights for both industry and academia.
- Developed a publicly available replication package for the study using Python and the Stack Exchange Data Explorer, enabling reproducible research, supporting future studies and further validation by the community.
- Collaborated with research teams to design experiments, build prototypes, and evaluate AI-based solutions for complex software engineering problems.
- Analyzed microservice antipatterns and compared the proposed decomposition approach with state-of-the-art methods. Used MicroFreshener and MicroMiner to detect antipatterns and assess the quality of generated microservice architectures in collaboration with the team.
- Built a simulator to process X-ray images using a novel method (SSRT) enabling non-invasive visualization of internal structures of 3D objects.
- Contributed to image processing advancements for non-destructive imaging, supporting applications in safe, harm-free object analysis.
- Rapidly mastered complex algorithms in the astra toolbox and 3D slicer, working effectively with research teams to meet publication objectives.
- Authored the paper "Performance Analysis of Computed Tomography Using the Scale-Space Radon Transform".
- Demonstrated strong problem-solving skills and adaptability in a cutting-edge research environment spanning medical imaging and computational tomography.
- Designed and developed custom AI solutions for clients across HR, logistics, and retail, addressing diverse business challenges.
- Built multi-agent AI systems using CrewAI and LangChain to automate business analysis and decision-making.
- Developed RAG-based chatbots and LLM-powered assistants by integrating OpenAI APIs, SQL databases, and n8n workflows.
- Created natural language-to-SQL solutions using LangChain SQL Agents, enabling non-technical users to query databases through conversational interfaces.
- Leveraged modern AI technologies, including LLMs, RAG, AI agents, vector databases, and workflow automation, to deliver scalable, production-ready applications.
- Designed and implemented a web application that facilitates the generation of web services for the national utility company, applying software engineering principles learned during academic studies.
- Organized and participated in academic events focused on artificial intelligence, enhancing community engagement and knowledge sharing among students and researchers.
- Created graphic designs including logos, charts, and scientific posters using Adobe Illustrator and XD to support AI-related events and research initiatives.
- Maintained club websites and supported digital initiatives for the scientific club at ESI, contributing to academic event promotion and community outreach.
Projects
AI solutions, research projects, and applications delivered across data science, and software engineering.
Microservices Research
Analyzed microservice antipatterns and compared the proposed decomposition approach BMSC-Decomp with state-of-the-art methods, including CoGCN, HierDecomp, and Mono2Micro. Used MicroFreshener and MicroMiner to detect antipatterns and assess the quality of generated microservice architectures in collaboration with the team.
Replication Package
Developed a replication package for the IEEE SANER 2025 study on microservices deployment trends, topics, and challenges in Stack Overflow using topic modeling with BERTopic. Also analyzed antipatterns in microservice applications and compared the proposed method with three existing approaches. The replication package enables reproducible research for the software engineering community.
Computed Tomography Simulator
Developed a CT simulator during a research internship using the novel Scale-Space Radon Transform (SSRT) for non-destructive X-ray imaging. Compared SSRT with the Radon Transform (RT) to evaluate image reconstruction quality.
RAG Chatbot with LangChain and Agent
An interactive chat system that answers questions based on specific documents and web pages while considering chat history. Prompts are structured using LangChain chains with memory, and the model was evaluated through both manual evaluation and LLM-assisted evaluation methods before integration into an autonomous agent.
Business Intelligence Agent
An intelligent BI system that converts natural language questions into SQL queries, retrieves data from databases, generates interactive Plotly visualizations, and provides AI-driven insights to support data-driven research and analysis without requiring SQL expertise.
AI-Powered Profiles Evaluation Agent
AI agent that automatically retrieves profiles from a database, ranks them based on keywords and positions, evaluates their suitability using OpenAI LLMs, assigns scores, and stores results in the database. The entire workflow is orchestrated end-to-end using n8n automation.
Crop Detector
Built an AI model and web application that detects the most suitable crop to plant on a defined soil, combining a React frontend with a TensorFlow/Keras backend trained on soil feature data.
ML Classification Models
Built and compared multiple ML classifiers: a stress detection model using physiological data (stress degree, heartbeats) served via Flask, and a Decision Tree / Naive Bayes classifier for categorical weather prediction using scikit-learn and graphviz.
Multi-Agent Analysis System
Built a suite of AI assistants using CrewAI and LangChain where specialized agents collaborate to analyze income, expenses, and investments, generating structured insights. Applied multi-agent coordination patterns relevant to AI research on agent collaboration.
AI Agent for Supplier Selection
An AI agent that selects the most suitable company for sourcing raw materials based on specific criteria provided in a DataFrame. The agent uses predefined tools to make decisions, with all actions and requests tracked and monitored through the Phoenix platform for evaluation purposes.
Arabic
French
English