Danny Gross

About Danny Gross

Strategist, systems thinker, and creative technologist building GenAI and agentic experiences with real-world impact at scale.At Meta Superintelligence Labs, I work across multi-agent systems, product behavior design, and evaluation infrastructure for personalized AI experiences. My focus is 0→1 work: standing up new capabilities, defining what \"good\" looks like before patterns exist, and building the evaluation frameworks to measure it.I also build internal tooling, creating infrastructure that helps teams work faster and measure better. I genuinely enjoy helping people learn how to work with new technology, so I make it a point to build and lead workshops, training sessions, and shared resources wherever I am.10+ years of storytelling, content design, strategic communications, and media systems experience has been foundational to how I work in AI. Context design, agent behavior, cross-functional translation, and knowing what high quality actually looks, feels, and sounds like all flow from that experience, combined with the natural language programming skills and systems thinking I\'ve built on top of it.

Experience

1. Content Engineer, Genai

   Meta

   Jul 2025 — Present · Seattle, WA, US

   I design agent and model behaviors, evaluation systems, and internal tooling that shape how AI products work in production, bridging product vision and model capabilities through cross-functional partnership on 0→1 AI product development.• Own agent behavior design for a personalized shopping agent launching into public testing - defining skills architecture, modular instruction frameworks, and conversation posture. Drove redesign from a monolithic system prompt to a modular skills-based architecture, proving the approach with a 70-30% pairwise win rate over production.• Led prompt design for a 9-agent content system in public testing with ~340K users and now a key growth lever for Meta AI - spanning behavioral design, tool calling, and specialized sub-agents• Developed and shipped end-to-end evaluation frameworks including a 4-dimension Information Quality Framework adopted into Meta AI Model Spec as company-wide standards and deployed as 6+ automated judges• Developed and shipped a prompt taxonomy for the shopping agent that maps user language and intent to evaluation criteria - enabling higher quality auto-judge evaluations and clearer baselines for improvement.• Built and shipped a custom MCP paired with a Claude Code skill - enabling evaluation development through natural language programming. Early adoption reduced a 5-hour workflow to 45 minutes.• Created internal tooling infrastructure to solve problems no existing tools addressed: a prompt version control platform, a content engineering performance dashboard for the VP of Design, and an auto-judge agent to measure skill triggering rate and accuracy based on user intent.

Education

• University of Washington

  Master's Degree, Communication, emphasis in Digital Media

  2015 — 2018

• West Texas A&M University

  Bachelor of Science (B.S.), Mass Communication; Electronic Media

  2006 — 2011

Skills

• Communications
• Video Editing
• Adobe Creative Suite
• Storytelling
• Editing
• Windows
• Project Management
• Avisynth
• Content Management
• Pro Tools
• Social Media Marketing
• Audio Engineering
• New Media
• Television
• Encoding
• Final Cut Pro
• Strategic Communications
• Photoshop
• Video Compression
• Digital Media
• Microsoft Office
• Teaching
• Photography
• Audio Editing
• Tutoring
• Leadership
• Avid
• Illustrator
• Social Networking
• Social Media
• Os X