Riccardo Aita

Experienced AI Data Trainer and Language Specialist with a strong background in prompt and response writing, chatbot evaluation, and multimodal data annotation, combined with proven expertise in AI Safety, Red Teaming, and Quality Assurance. Recently led multilingual teams in high-risk voice AI projects, overseeing evaluation workflows, conducting structured quality audits, and providing calibration and feedback to ensure alignment with safety policies and annotation standards. Specialized in high-latency voice model evaluation and conversational flow optimization, bridging the gap between text-based reasoning and natural vocal interaction. Skilled in assessing and auditing model outputs across dimensions such as instruction-following, truthfulness, tone, clarity, and safety, using methodologies including RLHF and SFT, and experienced in designing evaluation frameworks to improve model performance and consistency. Proven track record across multiple platforms, including audio transcription, video evaluation, and language-specific projects in Italian and English, with strong expertise in translating and reviewing English-to-Italian prompts and responses, including sensitive content, ensuring grammatical accuracy, fluency, and semantic fidelity. Passionate about advancing robust, safe, and high-quality AI systems through rigorous evaluation, cross-functional collaboration, and continuous improvement of model alignment and safety standards.