Amor Fati: Stress-Testing AI Video Systems by Breaking It on Purpose
A surreal short generated entirely through Google Flow, using one photo—my LinkedIn profile—as the only identity reference across wildly different aesthetics (cyberpunk, classical meets sci-fi, horror, video game).
The project probes system limits: identity drift, anatomical inconsistency, interpolation breakdowns, export artifacts, and UX friction at scale. Rather than optimizing for output, the work surfaces design insights for professional AI video workflows.
The result is both a personal narrative and a product-level case study in how AI video evolves from clip generator to creative infrastructure.
AIGC Flywheel for TikTok Shop E-Commerce
This multi-part project is a deep dive into how AI-generated content (AIGC) can accelerate GMV and build a sustainable growth flywheel for TikTok Shop.
I started from a human-centered design perspective—mapping seller pain points, buyer psychology, and creator incentives across the platform's stakeholders to define where AIGC delivers real value. I then analyzed TikTok Shop's two core content ecosystems—Shoppable Short-Form Video and Live Commerce—each with distinct conversion funnels and trust dynamics.
From this foundation, I developed a quality framework for AIGC covering content realism, category-specific standards, and data annotation strategies, along with UX principles for building user trust in AI-generated content.
Through workflow prototypes, I demonstrated how a single product can scale into region-specific, culturally adapted demos across APAC, Western, and Latin American markets—pushing toward content that feels authentic and creator-led.
This work represents my vision for how AIGC systems should be designed to serve commerce at scale.
*This project is password protected. Please reach out for access.