Loading...
Loading latest articles...
Please wait while we fetch the most recent tutorials.
Latest articles
Recent tutorials
Long-form guides that blend theory with practice. Each one includes math breakdowns, Python code, and links to the full notebook.
What you'll find here
Explore the site
Blog
In-depth tutorials on NLP, computer vision, forecasting, and responsible AI with diagrams and runnable code.
Read the articlesProjects
Hands-on builds with datasets, architectures, and Colab launchers you can run immediately.
See the projectsConsulting
Need help with a machine learning project? Let's discuss how we can support your team.
Get in touchFeatured projects
Notebooks on GitHub
Each project links to a GitHub repository and Google Colab notebook you can run immediately.
Reinforcement learning
Q-learning Tic-Tac-Toe bot and reward-tuning walkthroughs.
Open notebooksComputer vision
SRGAN super-resolution and OpenCV face detection.
Explore buildsTime-series forecasting
S&P 500 models and demographic forecasts with ARIMA/LSTM.
View forecastsNLP experiments
Multilingual T5 summarizer and Wikipedia-trained chatbot.
View NLP projectsAbout this site
What is Kudos AI?
Kudos AI is an independent educational platform created by Anas HAMOUTNI, a statistical engineer with a background in actuarial science, quantitative finance, and applied mathematics. The site is built around one idea: the best way to understand AI is to combine rigorous theory with hands-on practice.
Every tutorial on this site starts from the mathematical foundations, loss functions, gradient updates, attention mechanisms and walks through them step by step before presenting working Python code. All notebooks are open-source on GitHub, and most include a one-click Google Colab launcher so you can run every experiment yourself without installing anything.
The blog covers a wide range of topics across natural language processing, computer vision, time-series forecasting, and responsible AI. Articles range from introductory guides (like gradient descent explained from scratch) to deep technical breakdowns of architectures like Transformers and training techniques like RLHF.
Beyond tutorials, Kudos AI hosts a growing collection of reproducible ML projects, from SRGAN image super-resolution to Q-learning game agents and a set of interactive browser games that teach classification, optimization, and linear algebra concepts through play. Whether you are a student, a self-taught practitioner, or a professional looking to deepen your understanding, you will find original, carefully written content here.