Nathan Patra

Trained and evaluated large language models across conversational, instructional, and reasoning-based tasks. Assessed outputs for accuracy, coherence, relevance, tone, and contextual understanding; identified failure patterns such as hallucinations, bias risks, and edge cases. Crafted and reviewed high-quality prompts and applied structured reasoning frameworks for complex, multi-step evaluation. Edited and refined AI-generated content for clarity and natural readability, providing actionable feedback to improve alignment and performance.

Led shoot teams across preparation, on-ground direction, and post-production workflows, balancing creative decisions with logistics. Worked as a primary client-facing point to align needs, timelines, and budgets. Wrote copy for pitch decks, landing pages, scripts, taglines, and campaign messaging; refined marketing materials and supported product positioning. Collaborated with designers/artists to sharpen digital and print outputs. Identified actors and story angles for campaign needs (including UAE-focused projects), and coordinated with producers to finalize talent; reached out to editors/journalists for coverage and supported music/audio adjustments for promotional videos.

Led and trained a team of 160 writers producing annotation data across video, image, audio, and natural language formats for AI model training (including Adobe Firefly). Oversaw daily workflows, performance tracking, and efficiency improvements for high-volume, fast-paced production. Focused on dataset accuracy, consistency, scalability, and quality control while delivering large quantities of content weekly. Served as a client-facing role for international communications and data-driven insights; designed workflow strategies, defined KRAs, and introduced improved line processes and reporting mechanisms. Transitioned into training and communications after recognizing a passion for instructional leadership.