Mohamed SeifData Scientist & ML Engineer

I'm Mohamed Seif, an Egyptian data scientist and ML engineer based in 6 October City, Giza, Egypt. I build production machine learning systems from quantile pricing models and demand forecasting to synthetic data generation and LLM-powered assistants with multi-provider fallback chains. I work in both Arabic and English.

I'm open to freelance work in data science and ML engineering — predictive modeling, time-series forecasting, data pipelines, and LLM/RAG applications. The best way to discuss a specific project is to email me at mohamedseif.a1@gmail.com.

My strength is taking ML from notebook to production, not just training a model, but engineering reproducible pipelines, explainability, and serving infrastructure around it. KARNA is a good example: a real platform with a quantile pricing engine, SHAP explainability, monitoring, and a bilingual AI assistant, all containerized and production-ready.

KARNA is a used-car pricing platform for the Egyptian market. I built two core services: an ML pricing engine using e XGBoost and LightGBM with three-quantile predictions (q10/q50/q90) for negotiation ranges, SHAP explainability, and a model registry with hot-reload; and a bilingual Arabic/English AI assistant on a multi-provider LLM fallback chain with fuzzy car-name matching. The full stack runs across 5 Docker services.

At Datalentech I work as a junior Data Scientist, but I can't talk about any projects cause I keep work-specific details confidential, but I'm happy to talk about the techniques/skills I use at a high level.

I build LLM applications with LangChain (LCEL), applying prompt engineering techniques like few-shot, chain-of-thought, and structured Pydantic outputs, and tool integration. I completed a Generative AI internship with NVIDIA & ITI and an NVIDIA DLI workshop on building LLM apps with prompt engineering, and I shipped a bilingual assistant in KARNA.

Python first: Ensample models, tree-based models, Scikit-learn for classical ML, TensorFlow/Keras for neural networks, and SHAP for explainability, with Optuna for tuning. For deployment I use FastAPI, Docker, and Streamlit, with PostgreSQL/Supabase for data. For LLM work I use LangChain with structured outputs and multi-provider fallback chains.

I hold a B.Sc. in Computer Science & AI from Helwan University (GPA 3.51/4.0, Excellent), graduated June 2026, my graduation project is KARNA (Used cars dynamic pricing engine). I'm also an IBM Certified Data Scientist through Egypt's DEPI initiative, hold an NVIDIA DLI certificate in LLM applications, certified python data analyst from Datacamp, and earned an Honorable Mention at the ECPC (ICPC regional) programming contest.

I've worked in GIS and geospatial data science processing satellite and multispectral imagery, in poultry, in the automotive domain through KARNA, and in synthetic data generation for multi-scenario forecasting. And I am ready to adabt quickly to any new domain or data type.

The best way is email: mohamedseif.a1@gmail.com. You can also find me on LinkedIn via the links on this site. Send me a short note about what you're working on and I'll get back to you.