Towards Data Science has launched ShipAI, a video showcase for real-world AI projects. Practitioners record screen-share walkthroughs of their work, which are reviewed and published to help other builders understand the project's purpose, construction, and lessons learned. Each entry features a 4-15 minute video, AI-generated key takeaways, a searchable transcript, and stack notes detailing models, infrastructure, and tools.

Projects are linked to GitHub, Hugging Face, or live demos, though TDS does not host the work itself. Every builder receives a profile page collecting all their shipped projects. ShipAI opens with over 30 published walkthroughs from founding builders, mostly TDS authors.

Highlighted projects include a retrieval experiment on RAG chunk sizes, an AI damage-report generator for warehouse operations, and a CV transformed into a playable video game. Submissions are accepted via a form, requiring a screen recording using tools like Loom or OBS. The editorial team ensures content prioritizes educational value over production polish.

ShipAI is live at towardsdatascience.com/shipai, inviting the community to explore current AI work and consider sharing their own projects. For submission questions, contact [email protected].