Independent Multimodal QA Specialist (AI Red Teaming and UI/UX QA)
ImageRed Teaming
In this independent QA and red-teaming role, I designed adversarial test scenarios to stress-test Vision-Language Models (VLMs) against UI and generative model vulnerabilities. My efforts focused on identifying edge-case rendering artifacts, evaluating model robustness, and exposing prompt-based vulnerabilities. Using Python automation, I systematically validated AI-generated visual outputs and curated labeled datasets for model improvement.
• Conducted adversarial prompting and prompt injection attacks to challenge model boundaries.
• Detected and annotated visual artifacts in AI-generated images.
• Used Python scripts for batch validation and data upload.
• Ensured rigorous documentation for model evaluation and test reproducibility.
In this independent QA and red-teaming role, I designed adversarial test scenarios to stress-test Vision-Language Models (VLMs) against UI and generative model vulnerabilities. My efforts focused on identifying edge-case rendering artifacts, evaluating model robustness, and exposing prompt-based vulnerabilities. Using Python automation, I systematically validated AI-generated visual outputs and curated labeled datasets for model improvement. • Conducted adversarial prompting and prompt injection attacks to challenge model boundaries. • Detected and annotated visual artifacts in AI-generated images. • Used Python scripts for batch validation and data upload. • Ensured rigorous documentation for model evaluation and test reproducibility.
2025 - Present