Sulthan Josiah

I've contributed 103+ hours of speech-to-speech AI evaluation for Scale AI via Outlier, assessing multilingual voice outputs for naturalness, utility, audio quality, and conversational dynamics.

I identify hallucinations, interruption failures, embodied-experience claims, and audio artifacts, then write structured English-language justifications to support AI training data pipelines. I qualified for the advanced Live S2S Arena track after passing the Aether S2S Quality Screener, Stereo Headphone Check, and Audio Discrimination Test.

As a freelance AI content creator, I turn raw product images into photorealistic commercial staging, virtual apparel, UGC, architectural visualizations, and social assets designed for Instagram, TikTok, and marketplace listings. I've also created Kai, a consistent AI virtual influencer across 50+ images, and designed 200+ AI-generated tarot card illustrations using a custom visual system.

Earlier, I built an integrated project-monitoring dashboard at MIND ID, tracking 10+ strategic mining projects through Earned Value Management. My engineering background combines analytical project work with practical AI evaluation and creative production.