Jose Ferreira

I am a native Brazilian Portuguese AI data specialist focused on speech, image, and ad-quality evaluation. I perform verbatim pt-BR transcription, assess ASR output using WER methodology, and classify substitution, insertion, and deletion errors against reference transcripts. I am experienced with guideline-driven annotation workflows involving disfluencies, punctuation, number normalization, non-speech events, regional accents, audio quality, and intent classification. I have evaluated 1,500–2,000 images for generative AI training datasets and completed production ad-quality and policy-compliance work under high-throughput accuracy requirements. I also support AI response, search-quality, and data-collection workflows. My technical background includes Google Vertex AI, ComfyUI image-generation pipelines, n8n automation, JSON, Python fundamentals, and prompt engineering. I work professionally in English, communicate clearly in writing, and am available remotely in the UTC-3 time zone with availability for PST-hours training.