AI Engineering student · Mansoura National University
I'm Ibrahim Amin — I build machine learning systems that hold up outside the notebook: early-warning models for Egyptian agriculture, pass/fail classifiers for students, price predictors for retail. GPA 3.91, one competition deck away from graduation.
I'm an Artificial Intelligence Engineering student at Mansoura National University's Faculty of Engineering, in the accelerated track, expecting to graduate in July 2027. Most of what I know how to do, I learned by shipping it — a diamond-pricing model, a student early-warning system, a protected-area vegetation model — rather than only reading about it.
My training runs through DEPI, the Digital Egypt Pioneers Initiative's Data Science & AI track designed by IBM, and NVIDIA's Deep Learning Institute, where I worked through generative AI and applied LLM workflows. Both showed up directly in later project work.
Right now I own the Early Warning System module of AI AgriVision, our six-person graduation project — a bilingual decision-support system for Egyptian and Arab agriculture — while picking up freelance data analysis and Power BI work on the side.
Every tool below has shipped in an actual project further down this page — nothing here is aspirational.
Six projects, one thread: clean the data honestly, pick the model that earns its complexity, and hand the result to someone who has to act on it.
Graduation project. I own the Tier-3 fusion layer that turns weather, soil and vegetation signals into an early-warning score across 31 Egyptian protected areas, built on the real ERA5 + Landsat vegetation-and-climate dataset (1983–2025).
Competition project pairing a Power BI descriptive dashboard with a predictive classifier. Caught target leakage in the source features early, then handled a 13:1 class imbalance before trusting any accuracy number.
Built with a DEPI team on a 53,940 × 10 dataset. Compared classical and ensemble regressors before settling on XGBoost, with enough feature engineering and outlier handling to keep the generalization gap small.
Power BI analysis of ~99K orders on the Olist platform. Traced a 6.77% late-delivery rate to carrier transit time and connected it directly to satisfaction: ratings fell from 4.29 on-time to 1.76 undelivered.
Cleaned 1,155 transactions across 16 months — missing values, duplicates, outliers — then engineered time-based features to compare 6 product categories across 3 regions and 7 payment methods.
SQL-driven analysis of the Chinook database covering sales performance, catalog efficiency and a commission-restructuring model, delivered as a 32-slide team presentation across four analytical domains.
Mansoura National University, Faculty of Engineering — accelerated program, GPA 3.91/4.00.
Six-person team, supervised by Dr. Mohamed Zaki and Eng. Mennatallah Abdo Ouf. Owns the Early Warning System module end to end.
Via ITI. Prompt engineering, RAG-augmented LLMs, and applied generative AI workflows.
Hybrid practical training in Python, SQL, data analysis and machine learning; shipped the lumiere pricing model with a team.
Designed by IBM · Digital Egypt Pioneers Initiative · Python, SQL, ML, MLflow.
Beginner-level, delivered via the Information Technology Institute (ITI).
Ran a 6-session C programming workshop for complete beginners.
Organized events and resources for the student branch.
Supported new engineering students through enrollment; recognized by the Engineering Programs Director.
Open to graduation-project collaborators, freelance data analysis work, and teams who need someone to check the data before they trust the dashboard.