This project consisted of two interconnected subprojects.
1️⃣ Story Generation
The first part focused on building a system capable of constructing short narratives — ranging from 30 seconds to 5 minutes — based on a user prompt. The main challenge was designing an agentic workflow that could autonomously create coherent, engaging, and contextually consistent stories from minimal user input.
2️⃣ Video Production
The second part involved transforming the generated story into a finished video. This required deriving image prompts from the narrative, generating visually consistent scenes (in terms of style, characters, environment, and setting), and animating them with smooth camera movements. Additionally, I developed a custom subtitle system with word-level highlighting, precisely synchronized to the spoken narration.
My Role
I was the all-rounder on this project — responsible for everything from concept to deployment. I took ownership of the product, designed and implemented the cloud infrastructure, and built all data pipelines and APIs. I engineered sophisticated agentic systems powered by Generative AI, ensuring the entire workflow ran smoothly and was fully scalable. I also applied Infrastructure-as-Code principles and managed multiple environments for development, testing, and production.
Together, these components form an end-to-end pipeline that automatically produces short videos from user ideas — from story creation to final rendering. You can try it out here!