🎓 IBM Machine Learning Professional Certificate Portfolio
Professional machine learning portfolio built while completing the IBM Machine Learning Professional Certificate. Each project has been expanded beyond the coursework requirements into a production-oriented machine learning workflow emphasizing reproducibility, engineering best practices, and model evaluation.
- Created: August 25, 2026
- Updated: November 10, 2026
- Progress: Complete
This repository contains professional, production‑ready machine learning pipelines derived from the IBM/Coursera Machine Learning Professional Certificate (Courses 1–6). Each academic assignment has been rebuilt into a modular, reproducible, engineering‑aligned workflow using Python (and optional R).
The repo includes:
- Exploratory Data Analysis (EDA)
- Supervised ML: Regression
- Supervised ML: Classification
- Unsupervised ML
- Deep Learning & Reinforcement Learning
- Capstone Project
All projects follow software engineering and machine learning best practices:
- Modular project structure
- Reproducible experimentation
- Version control with Git/GitHub
- Feature engineering pipelines
- Cross-validation workflows
- Hyperparameter optimization
- Model benchmarking
- Statistical performance evaluation
- Data leakage prevention
- Automated preprocessing pipelines
- Comprehensive documentation
This repository serves as a comprehensive record of my machine learning development journey while demonstrating practical skills in data preparation, exploratory analysis, supervised learning, unsupervised learning, deep learning, and model deployment.
The IBM Machine Learning Professional Certificate has been completed. Future enhancements will focus on deployment, explainability, MLOps, and productionization of selected projects.
Completed the IBM Machine Learning Professional Certificate, covering exploratory data analysis, supervised learning, unsupervised learning, deep learning, recommendation systems, and machine learning engineering best practices through six progressively advanced projects.
This repository contains six end-to-end machine learning projects completed as part of the IBM Machine Learning Professional Certificate. Each project was expanded beyond the coursework requirements into a reproducible, engineering-focused machine learning workflow.
- 6 completed machine learning projects
- Multiple real-world datasets
- Supervised, unsupervised, and deep learning workflows
- Recommendation systems and neural collaborative filtering
- Production-oriented project structure
- Project Timeline
- Overview
- Datasets & Citations
- Professional Pipeline Design
- Cross‑Language Implementation
- Professional Engineering Standards
- Future Work
- Credits
- Started: September 9, 2026
- Last Edited: September 28, 2026
- Progress: Complete
- Notes:
- All Kaggle salary datasets merged, normalized, deduplicated, and exported.
- Foundation for Projects 01–03.
(Regression & Classification)
- Started: September 9, 2026
- Last Edited: September 14, 2026
- Progress: Complete
- Datasets Used: Kaggle Data Science Job Salaries (multiple sources)
- Notes:
- Numeric + categorical summaries
- Regression target:
salary_in_usd - Classification targets: mean/median salary categories
- Class imbalance checks
- Leakage detection
- Feature importance visualizations
- Started: September 17, 2026
- Last Edited: September 23, 2026
- Progress: Complete
- Datasets Used: Kaggle Data Science Job Salaries
- Notes:
- Models: Linear Regression, Ridge, Lasso, Decision Tree, Random Forest, Gradient Boosting
- Target:
salary_in_usd - Full numeric pipeline with CV + hyperparameter tuning
- Started: September 28, 2026
- Last Edited: October 3, 2026
- Progress: Complete
- Datasets Used: Kaggle Data Science Job Salaries
- Notes:
- Targets:
salary_mean_cat,salary_median_cat - Models: Logistic Regression, Random Forest, Gradient Boosting, SVM, KNN
- Full evaluation suite (accuracy, F1, ROC‑AUC, confusion matrix)
- Targets:
- Started: October 19, 2026
- Last Edited: October 30, 2026
- Progress: Complete
- Dataset Used: UCI Wine Quality Dataset
- Notes:
- Primary project: Wine Quality Unsupervised ML Pipeline
- Methods:
- Clustering: K‑Means, GMM, Agglomerative, DBSCAN
- Dimensionality Reduction: PCA, t‑SNE, UMAP
- Anomaly Detection: Isolation Forest, LOF
- Focus: latent structure discovery, cluster profiling, anomaly identification
- Will integrate with supervised pipelines for hybrid analysis
- Started: October 31, 2026
- Last Edited: November 4, 2026
- Progress: Complete
- Datasets: UCI Wine Quality Dataset
- Notes:
- Neural networks (dense, CNN, RNN)
- TensorFlow/Keras pipelines
- Reinforcement learning fundamentals (Q‑learning, policy gradients)
- GPU‑accelerated training workflows
- Will build on supervised + unsupervised pipelines
- Started: November 4, 2026
- Completed: November 10, 2026
- Progress: Complete
- Datasets:
- ratings.csv
- course_genres.csv
- course_embeddings.csv (generated)
- user_embeddings.csv (generated)
- Notes:
- Integrated concepts and techniques from Courses 1–5 into a comprehensive capstone project.
- Performed exploratory data analysis on 233,306 ratings from 33,901 learners across 126 courses.
- Developed content-based, clustering-based, and collaborative filtering recommendation systems.
- Implemented KNN, NMF, Neural Collaborative Filtering, Regression, and Classification models.
- Applied PCA, K-Means clustering, embedding learning, hyperparameter tuning, and model evaluation.
- Compared recommendation models using RMSE and identified Neural Collaborative Filtering as the top-performing approach.
- Demonstrated the complete machine learning lifecycle from data exploration through recommendation generation and model evaluation.
This repository contains professional, production‑ready machine learning projects based on the IBM Machine Learning Certificate:
- Exploratory Data Analysis
- Supervised ML: Regression
- Supervised ML: Classification
- Unsupervised ML
- Deep Learning & Reinforcement Learning
- Capstone Project
Each project is fully refactored into a professional pipeline emphasizing modularity, reproducibility, and engineering best practices.
Used for Exploratory Data Analysis, Supervised Machine Learning: Regression and Supervised Machine Learning: Classification Projects
This repo uses multiple publicly available Kaggle datasets:
-
Latest Data Science Job Salaries 2024 — Saurabh Badole
https://www.kaggle.com/datasets/saurabhbadole/latest-data-science-job-salaries-2024 -
Data Science Salaries 2023 — Arnab Chaki
https://www.kaggle.com/datasets/arnabchaki/data-science-salaries-2023 -
Data Science Salaries 2024 — Yusuf Delikkaya
https://www.kaggle.com/datasets/yusufdelikkaya/datascience-salaries-2024 -
Data Science Salaries and Fields — Josia Given
https://www.kaggle.com/datasets/josiagiven/data-science-salaries-and-fields -
Data Science Salaries — Sazid
https://www.kaggle.com/datasets/sazidthe1/data-science-salaries -
Data Science Salaries — Zain
https://www.kaggle.com/datasets/zain280/data-science-salaries -
Data Science Salary Landscape — Lai Nguy
https://www.kaggle.com/datasets/lainguyn123/data-science-salary-landscape(kaggle.com in Bing)
Cortez, P., Cerdeira, A., Almeida, F., Matos, T., & Reis, J. (2009). Wine Quality [Dataset].
UCI Machine Learning Repository.
https://doi.org/10.24432/C56S3T
- missing‑value profiling
- univariate & multivariate analysis
- salary distribution analysis
- geographic segmentation
- correlation heatmaps
- outlier detection
Models:
- Linear Regression
- Ridge / Lasso
- Decision Tree
- Random Forest
- Gradient Boosting
Engineering:
- CV
- hyperparameter tuning
- scaling
- pipeline automation
Models:
- Logistic Regression
- Decision Tree
- Random Forest
- Gradient Boosting
- SVM
- KNN
Evaluation:
- accuracy
- precision/recall
- F1
- ROC‑AUC
- confusion matrix
- clustering
- dimensionality reduction
- anomaly detection
- cluster profiling
- PCA/t‑SNE/UMAP visualization
- neural networks
- CNN/RNN architectures
- RL fundamentals
- TensorFlow/Keras pipelines
- exploratory data analysis
- content-based recommendation systems
- clustering-based recommendation systems
- collaborative filtering (KNN)
- matrix factorization (NMF)
- neural collaborative filtering
- embedding-based regression
- embedding-based classification
- recommendation model benchmarking
- end-to-end recommendation pipeline development
- Python
- R (optional)
- pandas
- NumPy
- matplotlib
- seaborn
- scikit-learn
- TensorFlow / Keras
- XGBoost
- imbalanced-learn
- PCA
- t-SNE
- UMAP
- K-Means
- Gaussian Mixture Models
- DBSCAN
- Agglomerative Clustering
- TensorFlow
- Keras
- Content-Based Filtering
- Collaborative Filtering
- K-Nearest Neighbors (KNN)
- Non-Negative Matrix Factorization (NMF)
- Neural Collaborative Filtering
- Embedding Learning
Throughout this repository I have implemented:
- End-to-end supervised learning pipelines
- Classification and regression workflows
- Ensemble learning methods
- Clustering and dimensionality reduction
- Neural network architectures
- Reinforcement learning foundations
- Automated hyperparameter tuning
- Feature engineering strategies
- Model evaluation and comparison frameworks
- Reproducible machine learning workflows
- Recommendation system design and evaluation
- Neural collaborative filtering
- Embedding-based feature learning
- Matrix factorization techniques
- modular scripts
- reproducible outputs
- version‑controlled notebooks
- standardized folder structure
- separation of data, models, evaluation
- production‑aligned pipeline design
- deploy recommendation systems using Streamlit
- add explainability techniques (SHAP, LIME)
- implement model monitoring and MLOps workflows
- experiment with transformer-based recommendation systems
- integrate cloud deployment solutions
- Based on coursework from: IBM Machine Learning Professional Certificate