This portfolio sample shows how I evaluate AI agents and automated workflows to find reliability problems, broken logic, and weak edge-case handling. I review whether the system completes the intended task, uses tools correctly, follows instructions, handles ambiguity, and produces an output that can actually be trusted.

The work focuses on identifying failure points before they become real user problems, including incomplete task execution, incorrect assumptions, workflow dead ends, hallucinations, and verification gaps. This is a portfolio demonstration of my AI agent evaluation and red-teaming work. No confidential client information is included.

Thorough breakdown! Robust edge-case handling and reliability checks are definitely the trickiest part of building reliable AI agents 🤖📊

Title: The Zen of Not Giving a Damn

WHAT I BUILT: A continuous 30-second photorealistic one-take sci-fi comedy film created natively with Seedance 2.5 in CapCut Video Studio.

The story follows Daniel, an unbothered office worker walking to work with his morning coffee while a full-scale apocalyptic alien invasion erupts around him—collapsing skyscrapers, airborne burning taxis, dimensional rifts, and colossal robotic walkers. Without a single cut, the camera tracks Daniel through revolving glass doors and security turnstiles into a serene open-space office, culminating in the ultimate workplace horror: a jammed printer while a 60-foot leviathan roars right outside the panoramic window.

WHY I MADE IT: To test the full cinematic limits of Seedance 2.5’s native 30-second continuous generation. My goal was to execute a seamless spatial transition (from extreme outdoor warzone destruction to structured corporate interior), consistent deadpan character acting, and photorealistic physics in one uninterrupted shot.

WORKFLOW: • Platform: CapCut Video Studio • Model: Seedance 2.5 (Single 30-second continuous one-take generation)

#contraseedancechallenge