Zachariah Sullivan

About Zachariah Sullivan

I design and build agentic AI systems that help companies use their internal documents, knowledge, and data to automate processes, improve decision-making, and scale operations. My work focuses on creating custom GraphRAG chatbots, internal workflow automation, and AI-driven marketing systems that turn company information into personalized content and higher conversion.With over 14 years of experience in growth marketing and data-driven strategy, I combine AI engineering with a strong track record of reducing acquisition costs, improving LTV, and building scalable programs. Previously, as Head of User Acquisition at Betfred Sportsbook, I led initiatives that reduced CAC and CPA by 75 percent, increased player value by 150 percent, and contributed to the company’s first profitable months. I also scaled multichannel campaigns across Google, Meta, Apple, X, and Instagram and improved ROAS through structured creative testing and data analysis.Today I work as a contractor building custom AI systems for companies. This includes developing agentic AI pipelines that organize and reason over internal data, implementing GraphRAG architectures for accurate and context-aware chatbots, automating manual workflows, and designing AI-powered marketing programs. My focus is helping businesses modernize operations, improve efficiency, and generate compounding ROI through well-designed AI systems.If your company is exploring how to integrate agentic AI or needs a custom solution built, I can help.

Experience

1. Ai Engineer

   3GIMBALS

   Nov 2024 — Present

   At 3GIMBALS, I worked as an AI Engineer and Statistical Programmer focused on building large-scale behavioral analysis systems for national security environments. I developed a risk taxonomy based on DoD guidelines that enabled structured, repeatable identification of high-priority risks within complex digital networks using digital exhaust and behavioral signals.I engineered agentic AI pipelines that automated information collection, data cleaning, and transformation, and pushed structured data into a modular knowledge-graph platform. This system powered a GraphRAG architecture with an internal chatbot that helped analysts surface hidden patterns, detect clusters of risky networks, and accelerate decision-making.The platform was designed to be scalable and modular, giving analysts a superintelligent assistant capable of synthesizing large volumes of data, reasoning over network relationships, and improving both the speed and quality of analytical assessments.

Education

• University of San Diego

  Bachelor of Arts (B.A.), Economics

  2010 — 2013

• UNLV Lee Business School

  Master’s Degree, Economics

  2015 — 2017