Advanced Technology
Machine Learning (ML)
Build models that learn from data for prediction, classification, and decision-making.
Prerequisites
No prior experience required — open to beginners.
Certificate
Awarded by Rescue Academy on successful completion of the program's assessments and final project.
Learn Online — Live Classes
Register, receive your schedule, and join live instructor-led Machine Learning (ML) classes on Zoom. Assignments, instructor feedback, and a certificate on completion.
Learn In Person
Attend Machine Learning (ML) classes in person in Juba with hands-on labs, instructor mentorship, and a certificate on completion.
What You'll Learn
- The machine learning workflow, data preparation and data leakage
- Regression, classification and model evaluation
- Feature engineering, pipelines, cross-validation and tuning
- Clustering, dimensionality reduction and imbalanced data
- Ensembles, interpretability and fairness auditing
- Time-series forecasting with backtesting
- Deploying, monitoring and governing ML models
Program Objectives
- Frame problems as ML tasks only where ML is appropriate
- Prepare data and build models without leakage, compared with honest baselines
- Evaluate, tune and interpret models in terms of real decisions
- Audit models for fairness and document them responsibly
- Deploy and monitor models as reliable services
Who It's For
- Python programmers who want to build predictive models
- Data and M&E officers who want to move from reporting to prediction
- Graduates preparing for data science and ML roles
Where You'll Use It
- Food security, health and education programmes (forecasting and risk scoring)
- Finance and mobile money (credit, liquidity and fraud models)
- Agriculture, logistics and business operations
Curriculum
Instructor-led, practical training. Each level is a block of 4, 8 or 12 weeks; every module ends with a practical activity and an assessment, and every level with a project.
Module 1: Machine Learning Foundations 10 h
Understand what machine learning is, the kinds of problems it solves, the steps of an ML project and the tools used in this program. 3 lessons
Module 2: Preparing Data for Machine Learning 11 h
Explore data, clean it, encode categories and split it correctly so that evaluation is honest. 3 lessons
Module 3: First Models: Regression and Classification 10 h
Train and interpret linear regression, logistic regression, k-nearest neighbours and decision trees. 3 lessons
Module 4: Evaluating Machine Learning Models 11 h
Measure model performance with the right metrics for the decision, and detect and reduce overfitting. 3 lessons
Module 5: Responsible ML and a First Project 10 h
Check data quality and bias, explain models to non-experts and complete the level project. 3 lessons
Level project: Predicting School Dropout Risk (training data)
Module 1: Feature Engineering and Pipelines 16 h
Create informative features, transform different column types correctly and combine every step into a scikit-learn pipeline. 3 lessons
Module 2: Cross-Validation and Model Selection 15 h
Estimate performance reliably with cross-validation, compare models against baselines and choose metrics that reflect real costs. 3 lessons
Module 3: Hyperparameter Tuning 13 h
Search for good model settings efficiently and honestly, and use learning curves to decide whether more data or a different model would help. 3 lessons
Module 4: Unsupervised Learning 15 h
Find groups and structure in unlabelled data with clustering and dimensionality reduction, and turn them into useful insights. 3 lessons
Module 5: Imbalanced and Messy Real-World Data 14 h
Handle rare outcomes, missing data and changing conditions that real projects always bring. 3 lessons
Module 6: Intermediate Machine Learning Project 17 h
Deliver the agent liquidity forecast and segmentation project end to end and present it to a business audience. 3 lessons
Level project: Mobile-Money Agent Liquidity Forecast and Segmentation (training data)
Module 1: Ensemble Methods 22 h
Combine many models with bagging, boosting and stacking for strong performance, and interpret what they learn. 4 lessons
Module 2: Interpretability, Fairness and Documentation 22 h
Explain individual predictions, measure and reduce unfair outcomes and document models so they are used responsibly. 4 lessons
Module 3: Time Series and Forecasting 22 h
Forecast quantities over time with baselines, feature-based models and validation that respects time order. 4 lessons
Module 4: Deploying Machine Learning Models 24 h
Package models reproducibly and serve predictions through APIs or batch jobs that others can run and trust. 4 lessons
Module 5: MLOps and Monitoring 20 h
Track experiments, version data and models, monitor production behaviour and govern retraining. 4 lessons
Module 6: Machine Learning Capstone 30 h
Design, build, deploy, monitor and document the Market Price Early-Warning Service as a professional ML project. 5 lessons
Level project: Market Price Early-Warning Service
Final practical project
Market Price Early-Warning Service
A deployed service that forecasts staple food prices for several markets, explains its forecasts with uncertainty, is monitored for drift and is documented with a model card.
Deliverables
- Backtested forecasting models compared with seasonal baselines
- A packaged forecasting service (API and monthly batch report)
- Experiment tracking, monitoring and a retraining runbook
- A model card readable by non-specialists
Assessed on
- Forecasts beat baselines in honest, per-market backtests
- The service runs from the README and validates inputs
- Monitoring detects a simulated drift and the response is documented
- Limits and appropriate use are clearly documented
Worked Examples
A selection of the visual, practical examples used in this program's lessons — each teaches a different skill.
- Workflow The life cycle of a machine learning project From the decision to support to a monitored model — and back again when the world changes. Beginner · Module 1 · Lesson 2
- Troubleshooting tree My model scores 99% — is it real? A checklist for the most common reasons a suspiciously good result is wrong. Beginner · Module 2 · Lesson 3
- Code example k-nearest neighbours from scratch: which crop for this plot? A ten-line k-NN that recommends a crop by finding the three most similar demonstration plots. Beginner · Module 3 · Lesson 2
- Worked calculation Reading a confusion matrix for a malaria screening model Accuracy, precision, recall and F1 calculated step by step from 200 test cases. Beginner · Module 4 · Lesson 2
- Data and chart When does a deeper tree stop helping? Training and validation accuracy as tree depth grows: the gap shows where memorising begins. Intermediate · Module 3 · Lesson 3
- Case study Auditing a microfinance credit model for fairness A composite, illustrative case: an accurate model that quietly failed rural women applicants, and how the lender responded. Advanced · Module 2 · Lesson 2
Free Course: Machine Learning Starter
A free introduction to machine learning: the main types of learning, the end-to-end workflow, a few key algorithms, evaluating models and the realities of data quality and deployment. Free, self-study, about 6 hours — no certificate. The full program above is taught live by an instructor, with practical assignments, assessments and a certificate.
- 1Free lesson 1What is Machine Learning? & Types of Learning
- 2Free lesson 2The Machine Learning Workflow
- 3Free lesson 3Key ML Algorithms & When to Use Them
- 4Free lesson 4Model Evaluation & Avoiding Overfitting
- 5Free lesson 5Real-World ML: Ethics, Data Quality & Deployment
- Final checkQuick Quiz — Machine Learning Basics
Tools & Technologies
Primary Tools
- Python with Jupyter or Google Colab — The working environment for every notebook and project Practice: Exploring data and running experiments
- pandas and NumPy — Loading, cleaning and transforming data Practice: Data preparation and feature engineering
- scikit-learn — The main ML library used throughout the program Practice: Pipelines, models, cross-validation, tuning and clustering
Supporting Tools
- Matplotlib — Plots for exploration, evaluation and communication Practice: Learning curves, confusion matrices and forecasts
- Git — Versioning code and notebooks Practice: Reproducible project history
Professional / Industry Tools
- XGBoost or LightGBM — Gradient boosting libraries used in the Advanced level Practice: High-performing tabular models
- FastAPI and Docker — Serving and packaging models Practice: Deployable, reproducible ML services
- MLflow — Experiment tracking and model registry Practice: Traceable, governed models
Career Opportunities
- Data/AI analyst role at a bank, telecom, or NGO
- Research or further-study pathway in AI/ML/Data Science
- Remote/freelance AI or data work
Practical Projects
- Train and evaluate a machine learning model on a real or public dataset
- Build a simple prediction tool and document its accuracy and limitations
- Given an imbalanced dataset, train a model and show why accuracy alone is misleading — report precision, recall, and F1 instead
Problem-Solving Focus
Problem: A model scores 99% accuracy on training data but performs poorly on new, unseen data
Approach: Recognize this as overfitting — use train/validation/test splits and cross-validation to catch it early, then simplify the model or add regularization instead of chasing higher training accuracy.
Problem: A dataset has far more examples of one outcome than another (e.g. 95% "no default" vs 5% "default" in a loan dataset)
Approach: Recognize that accuracy is misleading here — predicting "no" every time would still score 95% — and use precision, recall, and resampling techniques suited to imbalanced data instead.
Problem: A trained model needs to actually be usable by a real application, not just sit inside a notebook
Approach: Save the trained model and serve it behind a simple API endpoint that another application can call, rather than leaving it as a one-off script only you can run.
Problem: A model trained on historical hiring or lending data ends up replicating past unfair patterns against a particular group
Approach: Audit the model's predictions across different groups for fairness before deployment — overall accuracy alone can hide exactly this kind of bias.
Ready to register for Machine Learning (ML)?
Register — Learn Online Register — Learn In Person WhatsApp: +211926196668 Email: rescueacademy26@gmail.com