Data Science & Applied AI

Every Nigerian industry now drowns in data it cannot use: banks with millions of transactions, telcos with call records, hospitals with patient files, retailers with point-of-sale history. This nine-month track trains you to turn that raw material into decisions — and unlike most data programmes, it goes beyond notebooks into production: deploying, monitoring and maintaining models where real users depend on them.

The first third builds genuine foundations — Python, statistics, probability and the unglamorous craft of data wrangling, because 70% of real data work is cleaning inconsistent, incomplete records. Machine learning follows with classical algorithms taught properly (regression, trees, ensembles, evaluation) before deep learning with PyTorch. Then the applied AI module covers what the market is actually hiring for right now: working with large language models, building retrieval-augmented generation pipelines, prompt engineering, agents, and evaluating AI systems honestly.

The distinguishing final third is MLOps and your capstone. You version datasets and models, track experiments with MLflow, deploy a model behind an API, set up monitoring for drift, and ship a complete end-to-end product — past capstones include credit-scoring prototypes, churn prediction for telecom data and document-intelligence tools for Nigerian-language text. You present to a panel of practising data scientists and leave with proof you can carry ML from idea to production, which is precisely what separates hired candidates from certificate holders.