I turn machine learning ideas into usable software—from experimentation and model evaluation to reliable APIs and application integration.

Portfolio · LinkedIn · GitHub

I'm an Informatics and Artificial Intelligence graduate focused on building complete AI solutions that connect models, backend services, and real user needs.

• Build applications using LLMs, NLP, recommendation systems, computer vision, and expert systems
• Develop AI and backend APIs with Python, FastAPI, Pydantic, and asynchronous processing
• Train, evaluate, and integrate machine learning and deep learning models
• Care about explainability, reliable APIs, automated testing, and production-oriented engineering
• Open to Junior AI Engineer, AI/ML, and Backend AI opportunities

A three-class computer vision project for distinguishing real, fully synthetic, and tampered images, with five recorded ConvNeXt-Base/Small experiments and a PyTorch inference script.

• Compared dataset sizes of 3,976, 50,000, and 121,000 images
• Best saved validation accuracy: 94.45%; the experiment report records 94.24% test accuracy for the largest run
• Includes training notebooks, evaluation outputs, inference dependencies, and a README explaining the research and the related Streamlit/FastAPI application
• Model weights are supplied separately; the application source is not included in this research repository

Python PyTorch torchvision ConvNeXt Computer Vision Transfer Learning

An evidence-backed, bilingual recommendation system for software, AI, and cybersecurity projects. It asks 6–10 adaptive questions and returns up to three eligible tools with explanations, side-by-side comparisons, starter guides, and what-if previews.

• 192 catalog tools across Analysis, Design, Implementation, and Testing
• 12 project-stage/domain pools with 672 source-backed questions
• Hard requirement filtering for capabilities such as offline operation, free plans, and open source
• Arabic and English interface, with local-first short-text classification and an optional Ollama fallback

Python FastAPI JavaScript Ollama CLIPS Adaptive Questionnaires

Four AI-powered features I built as part of a wider digital coaching platform:

• Voice Coach: conversational voice assistance for coaching sessions
• Image Task Verification: checks whether uploaded images support completion of a described task
• Career Expert System: recommends technical career paths through an explainable, rule-based interview
• Fault Diagnosis: converts technical issue descriptions into evidence-backed troubleshooting plans

Python FastAPI NLP Computer Vision LLM Applications Expert Systems

My two contributions to a wider water-quality project: a machine learning workflow for predicting water-quality values and a rule-based FastAPI service for diagnosing symptoms and possible causes.

Python Machine Learning Random Forest FastAPI Expert Systems

• AI & Machine Learning: LLM applications, RAG, NLP, computer vision, recommendation systems, expert systems, model training and evaluation
• Backend Engineering: Python, FastAPI, Node.js, Pydantic, REST APIs, asynchronous processing
• Engineering Tools: Git, GitHub, Docker, automated testing
• Cloud: Building practical knowledge of Microsoft Azure and cloud deployment

I'm strengthening my skills in production AI backends, LLM application architecture, automated evaluation and testing, Docker, and Microsoft Azure.

I'm open to opportunities where I can contribute to real AI products, learn from experienced teams, and grow as an AI Engineer.