A practical introduction to AI for developers
For people who already write some Python and want to train real models — not only watch a keynote. Eight modules, about 20–30 hours, online.
Course overview
Fundamentals, then you train models: classical machine learning, deep learning, NLP and computer vision, plus the ethical questions that come with shipped systems.
- Audience: individuals with basic Python
- Format: online, self-paced — recorded videos, exercises, quizzes
- Length: 8 modules, about 20–30 hours
- Price: $199, with reductions when available (as on the French page)
- Platform: Teachable or a similar LMS
Eight modules, four blocks
Intro + Python refresh, then ML basics
What AI covers, a short history, sector examples, then NumPy and Pandas. Supervised / unsupervised / reinforcement learning, train/test splits, overfitting, precision, recall, F1. Exercise: a linear regression on a provided set.
- Python, NumPy, Pandas
- Learning types
- A first regression
Data prep and the algorithms you will actually use
Missing values, encoding, scaling, splits. Then linear and logistic regression, trees and random forests, SVMs. Exercise: clean a Kaggle-style set and compare models.
- Cleaning and scaling
- Regression, trees, SVM
- Compare performance
Tools and pace
- Jupyter Notebook, TensorFlow or PyTorch, scikit-learn
- Each module: 20–30 minute video, notes, code samples, a practical task, a quiz
- Suggested pace: one module a week
- Completion certificate at the end
The balance is theory, practice and ethics — skills you can take back to a real codebase.
Deep learning and NLP
Perceptrons, activations, backprop, CNNs and RNNs. Then tokenisation, bag-of-words, TF-IDF, sentiment, embeddings. Exercises: MNIST with TensorFlow/PyTorch, sentiment on IMDB-style reviews.
- Neural nets and CNNs
- Text preprocessing
- MNIST and sentiment
Computer vision and ethics
Image classification, detection, segmentation; pretrained ResNet and YOLO. Then bias, fairness, explainability, privacy and governance. Exercise: a biased-hiring case.
- Classify and detect
- Pretrained models
- Bias and responsibility
Training desk: +33 6 50 01 24 45 · contact@cybernecs.com
