AI Developer | Machine Learning Engineer
AI Developer at Optima Software Solution and Programming Instructor at School++ building medical AI, computer vision, LLM agents, RAG assistants, data-analysis tools, and backend AI products.
About
I work across medical imaging, genomics, tabular machine learning, RAG, agentic workflows, backend APIs, and data applications. My portfolio combines research-style experimentation with deliverables that can be tested, shared, and used by real people.
Technical Skills
Featured Projects
AI Agents
Planner, researcher, summarizer, and critic workflow for source-grounded Markdown/PDF research reports.
Includes web retrieval, evaluation metrics, Docker support, and Streamlit UI.Data AI
Agentic analytics app for CSV uploads, SQL exploration, profiling, charts, natural-language questions, and report export.
Built with modular typed Python services, SQL safety checks, and pytest coverage.LLM Engineering
Document assistant for PDF/DOCX question answering with multilingual embeddings and source citations.
FastAPI endpoints, Qdrant vector search, Streamlit UI, Docker, and retrieval evaluation.Medical AI
Consensus-annotation pipeline and deep learning benchmark for LIDC-IDRI pulmonary nodule segmentation.
Compared U-Net, Attention U-Net, U-Net++, DeepLabV3+, and SegFormer.Computer Vision
YOLOv8-based experiments for CT nodule detection and medical imaging workflows.
Focused on real-time detection, CT preparation, and research deliverables.Radiology AI
Deep learning notebooks for multiclass brain stroke diagnosis using CT/MRI imaging data.
Covers preprocessing, model training, evaluation, and research documentation.Bioinformatics
Risk prediction, prognosis, tumor subtype classification, and gene-signature analysis using clinical and genomic datasets.
Worked with TCGA-BRCA, METABRIC, BCSC, GEO, and doctor-facing decision support outputs.Precision Medicine
Predictive modeling workflow for invasive breast cancer chemotherapy response using gene-expression and clinical metadata.
Included dataset integration, external validation, and manuscript-style reporting.Biomedical Signals
Patient-level seizure prediction pipeline using multimodal EEG, ECG, and EMG windows.
Used ResNet-LSTM modeling, augmentation, focal loss, and checkpoint-based training.Conversational AI
Arabic customer-service chatbot for menu search, order handling, reservations, delivery, cart calculation, and handoff.
Built as a practical Streamlit prototype with structured JSON knowledge.Backend AI Product
Healthcare backend services covering core API flows, access roles, database models, validation, and operational workflows.
Structured endpoints for patient, doctor, appointment, and platform operations.Oncology AI
Research workflows for hepatocellular carcinoma materials, CEUS imaging, automated reports, and QA exports.
Focused on client-ready medical research analysis and reporting.Medical AI
CT and X-ray based workflows for lung cancer screening, image preprocessing, model training, and clinical-style evaluation.
Built for medical AI experimentation while keeping datasets, client context, and research details private.Medical ML
Classical machine-learning workflows for structured clinical data, risk scoring, classification, and model comparison.
Focused on clean features, validation, reproducible notebooks, and explainable results.Computer Vision
Plant-disease computer-vision workflow for classifying apple leaf disease and segmenting affected regions.
Designed as an applied CV pipeline with training, evaluation, and visual explainability outputs.Computer Vision
Monitoring system for face detection, fire, smoking, phone-call detection, attendance, and operational alerts.
Built around real-time detection workflows for workplace and restaurant operations.Accessibility AI
Image-captioning project to describe scenes and objects for visually impaired users.
Combines computer vision and language generation for practical assistive technology.Medical Chatbot
Conversational assistant for pharmacy-style medical questions, structured responses, and user-friendly guidance flows.
Built as a practical chatbot workflow with safety-aware medical communication patterns.Data Science
Time-series and machine-learning workflow for Egyptian tax prediction using recurrent modeling approaches.
Included GRU-style forecasting experiments and data-preparation workflows.Geo AI
Prediction workflow using night-time satellite imagery as a signal for demand and infrastructure planning.
Bridges geospatial computer vision, feature extraction, and predictive analytics.Software Product
Platform for managing academies, subscriptions, educational content, student operations, and payment-style workflows.
Startup-style product work covering dashboards, operational flows, and scalable product structure.Startup Product
Ride-sharing concept for connecting people around shared routes, trip requests, and community mobility.
Designed around practical user flows, location-based matching, and startup MVP thinking.Management System
Academic operations system for organizing faculty workflows, records, administrative views, and structured data.
Built as a practical software-management project with clear operational modules.Medical AI
Selected freelance medical projects are shown as case-study summaries only. Client code, unpublished research, datasets, and paper-related details stay private.
Medical Imaging
Deep learning workflows for CT-based lung analysis, screening support, and radiology research tasks.
Private case studyBreast Imaging
Breast-cancer analysis experiments combining imaging pipelines, feature engineering, and genetics context.
Paper in progressNeuro AI
Medical image classification and preprocessing workflows for stroke-related decision-support prototypes.
Private case studyNeuro Signal AI
Signal-processing and sequence-modeling experiments for early seizure prediction research.
Research summary onlyOncology
Applied ML notebooks and analysis workflows for HCC-focused clinical data exploration.
Private case studyBioinformatics
Drug-response and cancer-analysis experiments for research-oriented medical AI delivery.
Paper in progressMore On GitHub
Experience & Education
School++ | Present
Teaching programming fundamentals, computational thinking, problem solving, and project-based technology skills in a model similar to iSchool-style learning programs.
Optima Software Solution | Present
Building applied AI prototypes, data workflows, ML features, backend AI services, and client-ready software deliverables.
National Telecommunication Institute | Dec 2024 - Feb 2025
Completed a 3-month machine learning program with a final score of 96%.
Orange Digital Center & Instant | Aug 2022 - Oct 2022
Completed 150 hours covering software engineering, mathematics, machine learning, Python, data science, and deep learning with a 90% grade.
Comments & Reviews
Share this page with clients so they can leave a short public review.
Reviews will appear here after clients submit them.
Waiting for first client reviewContact
Cairo, Egypt | kareem.ayman.ml.ai@gmail.com | WhatsApp: +20 102 584 9793