Logo Lanfrica

Kevinobote/From-Predictive-to-Generative-AI

Domaine:

natural language processing
Créateur:
Kev
Hôte:
A hands-on workshop demonstrating the paradigm shift from traditional Predictive AI to modern Generative AI through live coding, prompt engineering, and multilingual AI experiences. Presented at: UbuCon Africa 2026 (1 hr 45 min) Technical Level: Intermediate (mixed coding experience welcome) Platform: Google Colab, VS Code, or local Jupyter # From Predictive to Generative AI: A Hands-On Journey Through the AI Revolution ## Workshop Overview A hands-on workshop demonstrating the paradigm shift from traditional Predictive AI to modern Generative AI through live coding, prompt engineering, and multilingual AI experiences. **Presented at**: UbuCon Africa 2026 (1 hr 45 min) **Technical Level**: Intermediate (mixed coding experience welcome) **Platform**: Google Colab, VS Code, or local Jupyter ## File Structure ``` . ├── UbuCon-2026/ # UbuCon Africa 2026 materials │ ├── 01_Predictive_AI_Titanic.ipynb # Traditional ML pipeline │ ├── 02_Generative_AI_Demo.ipynb # Generative AI with Ollama & Gemini │ ├── 03_Generative_AI_Kiswahili.ipynb # AI in Kiswahili language │ ├── presentation.tex # Beamer slides (22 slides) │ ├── presentation.pdf # Compiled presentation │ └── .gitignore # LaTeX auxiliary exclusions ├── Workshop_Presenter_Notes.md # Presenter guide ├── requirements.txt # Python dependencies └── README.md # This file ``` ## Workshop Structure ### Highlights - **3 hands-on notebooks** — Predictive AI, Generative AI, and Kiswahili AI - **22-slide Beamer presentation** for structured delivery - **Ollama-first approach** — local AI as primary, Gemini as secondary - **Multilingual AI** — full Kiswahili language experience - **Extended hands-on time** — 45 minutes of participant experimentation ### Timeline | Phase | Time | Content | Notebook | |-------|------|---------|----------| | Opening & Presentation | 0-15 min | AI landscape, key concepts | `presentation.pdf` | | Live Demo 1: Predictive AI | 15-35 min | Titanic ML pipeline, limitations | `01_Predictive_AI_Titanic.ipynb` | | Live Demo 2: Generative AI | 35-6 …