This repository contains the companion code for the DeepLearning.AI AI Coding Workflows course. It includes the project specifications, Python dependencies, and custom implementer-agent configuration you’ll use to build a simple Python web app throughout the course.
🎓 Keep learning → Explore all DeepLearning.AI courses — taught by the people building the future of AI. Find your next one.
💻 Explore more course artifacts → Browse the DeepLearning.AI course artifacts repo to find notebooks, projects, and notes from other courses across the DeepLearning.AI library.
The simplest way to take this course is to follow the videos in order, using the same project throughout.
The files in specs/ provide a consistent brief for the Python web app you’ll build repeatedly using different combinations of coding agents, cloud models, and local models. After each build, you’ll review and test the result before returning the project to a clean state for the next workflow.
Additional files are introduced as the course progresses, including the custom OpenCode implementer configuration used in the local-model lessons. All prompts used in the course are collected in prompts.md, so you can copy them rather than typing them from the videos.
Note: the copy of
.opencode/agents/implementer.mdin this repo is the version used in Lesson 11. In Lesson 9 you'll create a simpler version of it yourself. If you're working directly in a clone of this repo, delete or move that file until you reach Lesson 11, where the course tells you to bring it back.
The course was recorded on a Mac, so the local-model references in the implementer configuration and prompts use Mac values: MLX model builds and the .venv/bin/python test command. On Windows or Linux, substitute your GGUF model ID from LM Studio and .venv\Scripts\python.
| Path | Purpose |
|---|---|
specs/mission.md |
Describes the app, its purpose, and the experience it should provide |
specs/tech-stack.md |
Records the required technologies and implementation constraints |
specs/roadmap.md |
Breaks the app into phases that coding agents can plan and implement |
.opencode/agents/implementer.md |
Defines the custom implementer subagent used in the local-model lessons |
prompts.md |
All prompts used in the course, lesson by lesson |
requirements.txt |
Lists the Python packages required by the project |
.gitignore |
Excludes virtual environments, Python cache files, local configuration, and editor files from Git |
- Python 3.11 or later
- Git
- An IDE or editor—the course uses PyCharm
- Claude Code
- OpenCode
- An OpenRouter account
- LM Studio