Hire the Best Data Scientists
in Egypt

Publications A Hybrid Mutual Information- LASSO-Genetic Algorithm Selection Approach for Classifying Breast Cancer Classification of breast cancer using microarray gene expression data: A survey

Are you struggling with training a model, building a model from scratch, or understanding, interpreting, and transforming data into a model? I'm Ziad, an experienced AI engineer and data scientist. I specialize in solving complex problems using AI and data analytics solutions. I help companies like yours address challenges such as unstructured data, unclear trends, and inefficient decision-making processes. I teach you how to rationalize, train, or rebuild your model by providing clear, actionable insights and tailored solutions. I've completed numerous projects in all of these areas. Not only that, I'll complete any project for you at half the price you offer.

Behind on admin? Drowning in emails, invoices, spreadsheets, or project paperwork? Let me take those tasks off your plate so you can focus on growing your business—not managing it. I'm Mohand, a proactive Virtual Assistant with 2+ years of experience supporting a U.S.-based electrical contracting company. I don't just complete tasks—I create systems, keep projects organized, and make sure nothing gets missed. Here's how I can help: • Manage invoices, pay applications, and change orders • Organize project documents, contracts, and bid requests • Download construction plans and extract key project information into organized spreadsheets • Communicate professionally with contractors, suppliers, and clients • Track deadlines, payments, and project progress • Handle data entry, email management, research, and administrative support I'm experienced with Microsoft Excel, Word, Google Workspace, Procore, and I quickly learn new software and workflows. Accuracy, organization, and clear communication are what I bring to every project. If you're looking for someone who is reliable, detail-oriented, and takes ownership of the work instead of waiting for instructions, you're in the right place. My goal is simple: make your business run smoother while giving you back valuable time. Let's build a long-term working relationship—I’m ready to help your business stay organized, efficient, and ahead of schedule.

I build machine learning systems that ship to real users, not just notebooks that look good in a slide deck. At Zetta Global I built a Gradient Boosting forecasting pipeline across 116 locations and 20,000+ records, reaching R² = 0.93 and cutting error 67% versus baseline. I engineered 56 candidate features, narrowed to 20 using averaged Random Forest and Extra Trees importance, and caught same-year leakage features that were quietly correlated 99%+ with the target before training. The pipeline now extends into driver risk scoring on a PostgreSQL warehouse with millions of rows across safety event, mileage, and employee tables. For deep learning, I trained three image-classification models for the Wadjet Egyptian Heritage platform. The hieroglyph model hits 98.2% accuracy across 171 Gardiner classes. The landmark model reaches 93.8% across 52 Egyptian sites. Both run client-side via ONNX Runtime Web, so user images never leave the browser. Wadjet ships as a FastAPI + SQLAlchemy backend with JWT auth, integrates Gemini and Cloudflare Workers AI for image generation and TTS, and is deployed on Hugging Face Spaces with Docker. NLP and analytics work I have shipped: • BERT sentiment inference over Amazon reviews, comparing model output against star ratings across four alignment categories. • Stroke prediction with logistic regression written from scratch (gradient descent), plus a scikit-learn KNN baseline, wrapped in a tkinter GUI a non-technical user can drive. • Higher-education risk analysis at NTI: surfaced a "high GPA, zero engagement" risk segment that standard GPA-only rules miss. What you get when we work together: • End-to-end ML: data cleaning, feature engineering, modeling, deployment • Real production stack: PostgreSQL, FastAPI, ONNX, Docker, Hugging Face Spaces • Honest scoping. I tell you what I do not know, then go find out. • Clean handover. Documented, reproducible, written for the next person. Currently studying AI at Kafr El Sheikh University (2023 to 2027) and Head of Programming at IEEE KSU. Cairo timezone (GMT+2), available 30+ hrs/week. Send me the dataset, the goal, or the broken model. I will reply within 24 hours with a concrete next step.

I’m a Data Scientist with over 3 years of experience. Since 2022, I have helped clients transform unclear business questions and inconsistent datasets into tested, documented, and usable data solutions. I can help you with: • Data cleaning, validation, exploratory analysis, and feature engineering • Time-series forecasting and demand prediction • Classification, regression, clustering, churn prediction, and anomaly detection • Customer segmentation, recommendation systems, and behavioral analytics • Statistical analysis, hypothesis testing, and A/B testing • Power BI and Tableau dashboards • RAG systems, document Q&A, LLM chatbots, and AI workflow automation • Model deployment using Flask, Streamlit, Docker, and cloud services Selected results: • Developed a forecasting pipeline using 824K+ operational records, achieving an MAE of 3.22 and delivering dashboard-ready predictions. • Built a telecom churn model using 115K+ customer records, achieving 89.2% recall and an 80.4% F1 score. • Analyzed an A/B test involving 90,189 users using both frequentist and Bayesian methods. • Built customer-facing AI and automation solutions using Python, OpenAI and Claude APIs, RAG, Pinecone, n8n, Google APIs, and GCP. My process begins with clearly defining the target, success metric, and business constraints. I then validate the data, check for leakage, establish realistic baselines, compare suitable approaches, and communicate the results honestly. You will receive more than a notebook: clean and reproducible code, documented assumptions, meaningful evaluation, clear visualizations, and practical outputs that can be used after the project ends. Core stack: Python, SQL, Pandas, NumPy, scikit-learn, XGBoost, CatBoost, TensorFlow, PyTorch, Power BI, Tableau, Flask, Streamlit, Docker, GCP, LangChain, Pinecone, FAISS, OpenAI APIs, Claude APIs, and n8n. I’m a Computer Science graduate currently pursuing a Master’s in Data Science at Cairo University. If you have a dataset, business problem, or AI idea that needs to become a reliable working solution, send me a message with your objective and a sample of the data.

AI/ML Engineer — Computer Vision • NLP • LLMs • Voice Agents I build production-ready AI that ships fast and works reliably. From real-time voice agents to RAG chatbots and computer-vision pipelines, I turn messy data and research (e.g., RF-DETR) into products that move metrics. --- What I do - Computer Vision: Object detection & tracking (YOLOv8/YOLO-World, RF-DETR, Detectron2), OCR, video analytics, BYTETrack/SORT, OpenCV - LLMs & Agents: RAG over private data, tool-calling agents, NL→SQL, fine-tuning (LoRA/QLoRA, Unsloth, OpenRLHF/KTO), eval & safety (LLaMA 3, Qwen 2, Mistral, OpenAI) - Voice & Realtime: LiveKit + Twilio assistants with Whisper/ASR and XTTS/TTS; intent routing, verification, scheduling, workflow automation - Backend & MLOps: Python, FastAPI, LangGraph/LangChain, vLLM/Ollama, Transformers, Docker/CI, vector DBs (FAISS/Pinecone), PostgreSQL/MySQL --- Selected outcomes - Cut call-center load by 60% with a real-time voice agent (LiveKit + Whisper + XTTS + vLLM) - 3rd place – QTec Hackathon: CV-based supervision & anomaly detection (YOLO-World + multi-object tracking) - Top-3 Kaggle (public comp): book-price model with MAE = 3.4 - RAG + LLM apps that boosted engagement by 30% and reduced server response by 40% - Fine-tuned medical LLMs on OCR-generated datasets; shipped a production NL→SQL agent --- How I work - Product-first: clear problem framing, measurable KPIs, tight feedback loops - Production-grade: clean, documented code; tests, observability, graceful fallbacks - Research-aware: I track and adopt the latest papers/frameworks (e.g., RF-DETR) when they deliver ROI—not hype --- Hire me for - CV apps (detection/tracking/OCR), safety & proctoring, video analytics - Chatbots & RAG over your docs/data; agents that call tools/APIs - Voice bots for support, verification, scheduling, and automation - LLM fine-tuning/alignment, evaluation, and cost/perf optimization Let’s turn your AI idea into a reliable, high-impact product. Message me to discuss your project.