MOD-01level: beginner
Foundations
$ cat topics.txt
- → Python for data
- → NumPy / pandas
- → Stats you actually use
- → Git & notebooks
// outcome: Ship a clean EDA notebook on a public dataset.
student@hbtu:~$ ls modules/
modules.roadmap
Six tracks. Each one is a small, teachable chunk with a shippable outcome — no lecture marathons.
MOD-02level: core
Machine Learning
$ cat topics.txt
- → Regression & classification
- → Trees / ensembles
- → Model evaluation
- → Feature engineering
// outcome: Kaggle-style project with a public write-up.
MOD-03level: intermediate
Deep Learning
$ cat topics.txt
- → PyTorch basics
- → CNNs & transfer learning
- → Sequence models
- → Training loops from scratch
// outcome: Train a small vision or sequence model end-to-end.
MOD-04level: intermediate
NLP / LLMs
$ cat topics.txt
- → Tokenization & embeddings
- → Transformers
- → Fine-tuning HF models
- → RAG patterns
// outcome: Ship a mini LLM app with a real use case.
MOD-05level: core
Analytics & Viz
$ cat topics.txt
- → Metrics that matter
- → Dashboarding
- → Storytelling with data
- → A/B basics
// outcome: One dashboard + one insight story you can defend.
MOD-06level: advanced
MLOps lite
$ cat topics.txt
- → Reproducible envs
- → Experiment tracking
- → Model serving basics
- → CI for notebooks
// outcome: Deploy a model behind an API you actually maintain.