# MUSE: Multi-Tenant Model Serving With Seamless Model Updates

DevFeed: [MUSE: Multi-Tenant Model Serving With Seamless Model Updates](<https://devfeed.tech/articles/muse-multi-tenant-model-serving-with-seamless-model-updates-126681.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2602.11776>)

Author: Cl\\'audio Correia, Alberto E. A. Ferreira, Lucas Martins, Miguel P. Bento, Sofia Guerreiro, Ricardo Ribeiro Pereira, Ana Sofia Gomes, Jacopo Bono, Hugo Ferreira, Pedro Bizarro

Published: 2026-11-10T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Distributed Computing](<https://devfeed.tech/sources/arxiv-distributed-computing.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [AI Integration Strategies](<https://devfeed.tech/topics/ai-integration-strategies.md>), [tenant data protection](<https://devfeed.tech/topics/tenant-data-protection.md>), [cost-optimization](<https://devfeed.tech/topics/cost-optimization.md>), [zero downtime deployment](<https://devfeed.tech/topics/zero-downtime-deployment.md>)

Tags: [cs-dc](<https://devfeed.tech/tags/cs-dc.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [fraud](<https://devfeed.tech/tags/fraud.md>), [high-availability](<https://devfeed.tech/tags/high-availability.md>), [lead-time](<https://devfeed.tech/tags/lead-time.md>), [low-latency](<https://devfeed.tech/tags/low-latency.md>), [model-serving](<https://devfeed.tech/tags/model-serving.md>), [multi-tenancy](<https://devfeed.tech/tags/multi-tenancy.md>), [muse](<https://devfeed.tech/tags/muse.md>), [resilience](<https://devfeed.tech/tags/resilience.md>), [routing](<https://devfeed.tech/tags/routing.md>), [sharing](<https://devfeed.tech/tags/sharing.md>), [stable](<https://devfeed.tech/tags/stable.md>), [transformation](<https://devfeed.tech/tags/transformation.md>)

## AI overview

MUSE addresses a multi-tenant serving bottleneck: retraining shifts model score distributions, making client-managed decision thresholds stale and costly to recalibrate. It shares models through dynamic intent-based routing and applies a two-level score transformation to map outputs to a stable reference distribution, decoupling serving updates from client decision boundaries. The abstract reports Feedzai deployment across several dozen tenants, processing over 1,000 events per second and more than 55 billion events in the last 12 months with high-availability and low-latency guarantees. It says model lead time fell from weeks to minutes; detailed implementation and evaluation methods are not included in the supplied excerpt.

## Source excerpt

arXiv:2602.11776v3 Announce Type: replace-cross Abstract: In binary classification systems, decision thresholds translate model scores into actions. Choosing suitable thresholds relies on the specific distribution of the underlying model scores but also on the specific business decisions of each client using that model. However, retraining models inevitably shifts score distributions, invalidating existing thresholds. In multi-tenant Score-as-a-Service environments, where decision boundaries reside in client-managed infrastructure, this creates a severe bottleneck: recalibration requires coordinating threshold updates across hundreds of clients, consuming excessive human hours and leading to model stagnation. We introduce MUSE, a model serving framework that enables seamless model updates by decoupling model scores from client decision boundaries. Designed for multi-tenancy, MUSE optimizes infrastructure re-use by sharing models via dynamic intent-based routing, combined with a two-level score transformation that maps model outputs to a stable, reference distribution. Deployed at scale by Feedzai, MUSE processes over a thousand events per second, and over 55 billion events in the last 12 months, across several dozens of tenants, while maintaining high-availability and low-latency guarantees. By reducing model lead time from weeks to minutes, MUSE promotes model resilience against shifting attacks, saving millions of dollars in fraud losses and operational costs.