David Mazzarri

Contributed to advanced AI training initiatives by developing system instructions, golden responses, and structured annotations to refine large language model writing capabilities. Designed and authored high-quality prompt-response pairs to train AI systems in narrative structure, tone calibration, clarity, and stylistic consistency. Led a specialized humor-focused training project, rewriting AI outputs to enhance wit, timing, and tonal nuance while defining contextual boundaries for appropriate humor use. Provided editorial oversight on model outputs, identifying linguistic inconsistencies and improving coherence, voice alignment, and user engagement. Applied structured evaluation frameworks to assess response quality, ensuring alignment with user intent, brand tone, and contextual sensitivity. Collaborated in iterative feedback cycles to improve model performance across creative, conversational, and instructional use cases.

Developed brand messaging and crafted 200 engaging, interchangeable messages for a novel line of trucker hats designed to foster connections among wearers. Contributed to product design by collaborating on material selection, aesthetic elements, and overall market positioning to enhance appeal and user engagement. Led the development of a 12-month multimedia campaign, creating over 100 assets—including video, audio, and social media content—optimized for all platforms to drive high engagement and brand visibility.

Developed brand messaging for a novel line of trucker hats and contributed to product design, including material selection and aesthetics. Led a 12-month multimedia campaign, producing over 100 assets across video, audio, and social platforms to boost engagement and brand visibility.