PEPR 26 - Responding to Regulations Confidently and Efficiently @UsenixOrg
PEPR 26 - Responding to Regulations Confidently and Efficiently  @UsenixOrg
Uploaded June 2026 | Updated September 2026, 3 weeks ago
Responding to Regulations Confidently and Efficiently

Marek Matejka, Meta

Regulatory requirements, particularly those mandating user setting changes, frequently pose a significant challenge due to their volume and potential impact, if implemented incorrectly. This talk presents a tried-and-tested process and engineering framework developed at Meta to respond to these regulations both confidently and efficiently. We detail the multi-step approach, beginning with critical cross-functional (XFN) alignment to define scope and setting values. We then provide an overview of the core framework, which drastically reduces implementation time—achieving code-completeness for setting-related requirements within a single working day. Finally, we discuss the complete lifecycle, emphasizing mechanisms for a confident launch, and, crucially, ongoing compliance. We explore the use of the Swiss Cheese Model for high assurance, robust monitoring, and solutions for complex edge cases such as conflicting regulations or regulatory overlap. This framework is a blueprint for scaling effective and trustworthy regulatory responses that others can adopt and build on top of.

View the full PEPR '26 program at usenix.org/conference/pepr26/program
PEPR 26 - Responding to Regulations Confidently and EfficientlyNSDI 26 - FlexLLM: Token-Level Co-Serving of LLM Inference and Finetuning with SLO GuaranteesNSDI 26 - Libra: Flexible Request Partitioning and Scheduling for Serving Unbalanced and Dynamic...SREcon26 Americas - The Ironies of AI²USENIX Security 25 - Serverless Functions Made Confidential and Efficient with Split ContainersNSDI 26 - Enabling AI Network Cross-Layer Design and Operations with Arcadia...NSDI 26 - KRAKENGUARD: Towards Fine-Grained eBPF IsolationNSDI 26 - Eywa: Automating Model-Based Testing using LLMsSREcon26 Americas - Executing Chaos Engineering in Production at a Critical Financial InstitutionSREcon26 Americas - The Case of the Misnamed Cities: CAST Analysis of a Google Maps IncidentNSDI 26 - Attack of the Bubbles: Straggler-Resilient Pipeline Parallelism for Large Model TrainingNSDI 26 - Matryoshka: Realizing Hyperscale Data Center Network Design for the AI Era
USENIX |

PEPR '26 - Responding to Regulations Confidently and Efficiently

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER