Deep Learning 2026

This repository contains the course material for the deep-learning part of the 2026 course.

The course is short and strongly practice-oriented. The goal is not to cover the whole field in mathematical depth. The goal is to help you:

• get a broad overview of modern deep learning
• choose a technical path
• choose an application domain
• build a unique AI application or prototype
• understand and defend what you built

Deep learning is a broad field. In this course, you might work on:

• computer vision
• language models and LLM apps
• multimodal and document AI
• semantic search and retrieval
• recommendation systems
• speech and audio
• time-series and sensor signals
• 3D, depth, and geometry
• reinforcement learning and control
• edge AI and deployment

How To Use This Repository

The simplest workflow is:

1. read 01-assignment-brief.md
2. read 02-technical-paths.md
3. read 03-application-domains.md and then explore one or more files in domains/
4. read 04-supporting-tools.md for supporting tools such as web interfaces, APIs, frontend choices, and packaging

Repository Structure

• 01-assignment-brief.md Assignment expectations and scope.
• 02-technical-paths.md High-level map of the main technical paths you can choose.
• 03-application-domains.md High-level map of the main application domains, with links to the matching files in domains/.
• 04-supporting-tools.md Optional shared tooling for interfaces, APIs, frontend choices, and packaging.
• domains/ One short file per student-facing project domain.
• slides/ Quarto-based slide deck source plus the generated PowerPoint for the kickoff presentation.

Teaching Philosophy

• broad rather than deep
• practical rather than purely theoretical
• build something real
• use AI tools if useful, but understand your own code

The slide deck is a shorter subset of the written notes. If you want more context after the presentation, start with the top-level Markdown files and then open the matching file in domains/.