About Zivis
We don’t just look for flaws in your code or infrastructure—we probe the behaviors, prompts, and blind spots of your AI systems. From autonomous agents to fine-tuned models, Zivis helps you understand what your AI might do when no one’s watching—and what a real adversary could make it do.
Why "Zivis"?
The name Zivis comes from the Latin civis, meaning “citizen.”
We chose it because AI doesn’t just live in systems—it lives in user interactions. In prompts. In decisions that affect people.
Our mission is to pressure-test AI the way it can be threatend—unexpectedly, imperfectly, maliciously, and adversarially.
Because every AI system, in the end, has to face the public.
Jake Miller
Founder & CEO
​
Jake Miller is a 23-year veteran of the software industry, known for building scalable platforms and leading engineering teams at the intersection of product innovation and security. He’s the founder of Zivis, a solo practice offering AI-augmented red teaming and vulnerability testing for teams building with LLMs, RAG, and agentic systems.
Before founding Zivis, Jake led Engineered Innovation Group, where he helped dozens of SaaS companies and IT teams design, build, and secure AI-driven products and platforms. His experience spans enterprise SaaS, startup acceleration, and large-scale system architecture—including his time as the first engineer and later Director of Engineering for Salesforce Marketing Cloud’s Journey Builder.
​
Jake is actively focused on building deep specialization in red teaming and AI security, bringing the same rigor and creativity to this discipline that he's brought to two decades of software engineering. At Zivis, he leads every engagement personally—combining human judgment, advanced automation, and emerging threat models to help teams adopt AI safely and confidently.