Private-by-Design AI for Video and Beyond
Private-by-Design AI for Video and Beyond
Connect video, docs, and transcripts with privacy preserved across the pipeline.
Wednesday, November 19, 2025 • 11:00 AM PT
Hosted by AI Collective | Featuring Protopia AI, Miriel AI, and CyborgDB
Most AI systems hallucinate because they lack context. Most are hard to deploy because they expose sensitive data. In this session, Miriel shows how a context engine grounds AI in real data sources like video, documents, and transcripts. Protopia and Cyborg show how to keep that data private from ingest to inference. You leave with a blueprint for building context-aware, privacy-first AI that works in production.
You will learn
• How a context engine reduces hallucination by unifying video, docs, and transcripts
• How to protect visual and textual data in use and at rest
• How encrypted vector search blocks data extraction attacks
• How to design an end-to-end privacy-first pipeline that still performs
Who should attend
Security and data leaders, AI architects, platform teams, compliance owners in finance, healthcare, public sector, and any company needing on-prem or private cloud options.
Speakers
Josh Paulson, Miriel AI. Josh is one of the founders of Miriel, the AI context engine for developers. He leads all functions outside of engineering and partners with teams to ship AI into production. With over a decade of experience building, scaling, and investing in startups, he's launched products used by millions. Josh was RStudio's second hire and spent years in VC and hedge funds. He holds the Chartered Financial Analyst (CFA) designation, is a Macalester grad, and can impersonate Mickey Mouse on demand.
Nicolas Dupont, CyborgDB. Nicolas is the CEO & Founder of Cyborg, a startup enabling AI adoption in regulated industries through Confidential AI. He leads the team's work on CyborgDB, the first end-to-end encrypted vector database which underpins secure AI inference workflows.
Jennifer Cwagenberg, Protopia AI. Jennifer leads engineering at Protopia AI, ensuring the Stained Glass Transform meets customer needs for data protection. One of the company’s earliest hires, she previously held engineering and security roles at Toyota, N-able, SolarWinds, and Match.com. With deep expertise in large-scale data and cloud security, Jennifer holds an MS and BS in Computer Science from the University of Texas at Dallas.
Moderator: Charlcye Mitchell, CyborgDB. Charlcye is Senior Product Manager at Cyborg, where she bridges advanced encryption architectures with real-world AI deployments in regulated industries. With 20+ years in technical leadership, she’s built and led global engineering teams and delivered products used by millions. Charlcye focuses on translating cutting-edge cryptography into deployable systems for federal and enterprise customers, making privacy-preserving AI practical for developers in healthcare and fintech.