Mariam Ayman

AI/ML Engineer

Building intelligent solutions that improve lives and drive smarter decisions.

View Projects Contact Me

About

Hi, I'm Mariam Ayman, an AI/ML Engineer. I'm early in my journey, but I'm driven by a clear conviction: AI isn't just about code — it's about building solutions that improve lives and drive smarter decisions.

Through hands-on projects and continuous learning, I've discovered my passion for solving real-world problems with machine learning. I don't just chase accuracy metrics; I think deeply about how models create genuine value and impact.

What sets me apart? I prioritize understanding concepts deeply rather than just applying libraries. I believe that true expertise comes from grasping the 'why' behind algorithms, not just the 'how.' This foundation allows me to build thoughtful, effective solutions that solve meaningful problems — and I'm excited to grow this expertise with every project I tackle.

My approach to solving problems:

  • • End-to-end ML development – From problem framing to deployment, building solutions that create measurable impact.
  • • Deep theoretical foundation – Understanding the 'why' behind algorithms, not just applying libraries blindly.
  • • Value-driven modeling – Focusing on business impact and real-world results, not just accuracy metrics.
  • • Clean, maintainable code – Writing production-ready code with best practices and version control.
  • • Continuous learning mindset – Staying current with latest techniques while mastering fundamentals.
"My focus is not only on achieving high accuracy rates, but on designing models that create genuine value and tangible impact."

Skills & Expertise

Programming Languages

  • • Python
  • • JavaScript
  • • HTML5 & CSS3
  • • C
  • • Java
  • • SQL

Data Analysis & ML

  • • Machine Learning
  • • Scikit-learn
  • • Pandas
  • • NumPy
  • • Statistical Analysis
  • • Feature Engineering

Data Visualization

  • • Data Visualization
  • • Interactive Dashboards
  • • Statistical Analysis
  • • Matplotlib
  • • Seaborn
  • • Plotly

Tools & Technologies

  • • SQL Databases
  • • Data Cleaning
  • • Data Engineering
  • • Git & GitHub
  • • Jupyter Notebooks
  • • VS Code

Projects

E-commerce Delivery Delay Prediction
Analyzed 100,756 Brazilian e-commerce orders to predict delivery delays. Performed data cleaning, feature engineering, and exploratory analysis on 96,478 delivered orders.
XGBoost: 92.7% Random Forest: 91.4%
Key insight: Shipping cost (37.6%), order value (34.0%), and payment installments (28.3%) are top predictors of delivery time.
Read Case Study
High-Value Customer Prediction
Analyzed 392,692 transactions to identify high-value customers. Engineered customer-level features including frequency, monetary value, and purchase patterns.
Random Forest: 90% PCA: 100% variance
Applied PCA (72.9% + 27.1% variance explained) and ensemble methods. UK leads sales, followed by Netherlands, Ireland, and Germany.
Read Case Study
Heart Disease Risk Prediction
Built classification models on 16,859 health records to predict heart disease risk. Key features: BMI, age group, sleep quality, physical health.
Logistic Regression: 91.1% Random Forest: 89.8%
Dataset split: 13,487 training / 3,372 testing. Identified critical health indicators for early intervention.
Read Case Study
Customer Churn Prediction
Analyzed 5,630 customer records to predict churn and understand behavior. Features include demographics, purchasing patterns, and engagement metrics.
XGBoost: 97% Precision: 95%
Key insight: Tenure, complaints, and inactivity are strongest churn predictors.
Read Case Study

Interested in working together?

Let's turn your data into actionable insights.

Let's Connect →

Contact

mariam.ayman62276889@gmail.com linkedin.com/in/mariam--ayman github.com/mariamayman540