Uploaded June 2026 | Updated November 2026, 3 days ago
AI agents are rapidly moving from demos to production, acting autonomously across tools, data systems, and workflows—and in the process, they amplify data movement far beyond what traditional governance models were designed to handle. Data security controls built for humans break down when agents operate at machine speed, execute in parallel, and persist sensitive information across new data surfaces like trajectories, embeddings, logs, and tool outputs.
In this talk, we outline the emerging data governance failures in agentic architectures—identity confusion for data access, entitlement creep, recursive leakage across agent chains, new data constructs leading to old controls becoming obsolete, and why out of box agent harnesses and existing IAM are insufficient. We then present Meta’s governance-first approach for safely enabling agents at scale: a defense-in-depth stack centered on Isolation Domains (domain-scoped encryption and output closure), Agent Identity (end-to-end attribution distinct from the user), Agent-Aware Access Control (classification-aware ABAC evaluated at query time), AccessMate (zero-standing-permissions access triage and least-privilege fallback), CodeGuard (secure code generation and runtime execution guardrails), and DataVM—a unified trusted data environment that bounds inputs, tools, and outputs under one governed scope.
Attendees will leave with a concrete reference architecture for building agents that are not merely powerful, but governable, auditable, and regulatory-ready—turning governance from a blocker into the harness that safely unlocks agent autonomy.
Want to learn more about the @Scale Conference? Visit our website here: atscaleconference.com
AI agents are rapidly moving from demos to production, acting autonomously across tools, data systems, and workflows—and in the process, they amplify data movement far beyond what traditional governance models were designed to handle. Data security controls built for humans break down when agents operate at machine speed, execute in parallel, and persist sensitive information across new data surfaces like trajectories, embeddings, logs, and tool outputs.
In this talk, we outline the emerging data governance failures in agentic architectures—identity confusion for data access, entitlement creep, recursive leakage across agent chains, new data constructs leading to old controls becoming obsolete, and why out of box agent harnesses and existing IAM are insufficient. We then present Meta’s governance-first approach for safely enabling agents at scale: a defense-in-depth stack centered on Isolation Domains (domain-scoped encryption and output closure), Agent Identity (end-to-end attribution distinct from the user), Agent-Aware Access Control (classification-aware ABAC evaluated at query time), AccessMate (zero-standing-permissions access triage and least-privilege fallback), CodeGuard (secure code generation and runtime execution guardrails), and DataVM—a unified trusted data environment that bounds inputs, tools, and outputs under one governed scope.
Attendees will leave with a concrete reference architecture for building agents that are not merely powerful, but governable, auditable, and regulatory-ready—turning governance from a blocker into the harness that safely unlocks agent autonomy.
Want to learn more about the @Scale Conference? Visit our website here: atscaleconference.com



![Spatial Media on Instagram VR | Jaime Del Palacio, Meta
Jamie Del Palacio, a software engineer on the Instagram VR team, delivered a presentation on the process of transforming the Instagram application into a true virtual reality experience on Quest headsets. [00:05]
He detailed the teams journey to convert all 2D media into 3D stereoscopic content using AI. [02:00]
Palacio outlined the primary challenges they faced, including the immense scale of Instagrams content, the critical need for low-latency processing, and ensuring user comfort. [04:03]
He explained their decision to convert media on-demand on the server and detailed the optimization strategies—such as parallelization and code tuning—that reduced conversion times from over two minutes to under ten seconds. [07:17]
The talk also covered quality considerations, like prioritizing a comfortable viewing experience (positive parallax) over more jarring 3D effects and developing a specific solution for rendering text overlays correctly. [12:28] He concluded by sharing key learnings and outlining future ambitions for creating even more immersive social experiences. [00:19] Spatial Media on Instagram VR | Jaime Del Palacio, Meta](https://i.ytimg.com/vi/da0ACS29mUQ/mqdefault.jpg)






