CNTUG Meetup 2025/08 SP #70

Name: CNTUG Meetup 2025/08 SP #70
Start: 2025-08-14T19:30:00+08:00
End: 2025-08-14T21:30:00+08:00
Location: 五倍學院 - 實戰轉職・企業內訓．技術諮詢服務

Cloud Native Taiwan User Group

Aug 14, 11:30 AM – 1:30 PM (UTC)

In-person event

About this event

（本活動於 2025/7/31 12:00 PM (UTC+8) 開放 RSVP，需收 50 元場地費，歡迎大家來一同參與！）
哈囉大家好，8/14 (四) 我們再次來到「五倍學院」舉辦「CNTUG 2025/08 Meetup SP #70」。欸？我們不是有 8 月 COSCUP 的 meetup 嗎？確實是，這場是加開的八月 meetup，因為我們特別從遠方來的朋友分享議程！
首先第一場 session，由 Hrittik Roy 帶來「Multitenancy in the Kubernetes Era」，他目前是 CNCF Ambassador 也是 Loft Labs 的  Platform Advocate。
多租戶（Multitenancy）是一種讓多個團隊共享同一個 Kubernetes 叢集的實作方式，能有效提升資源使用率、降低營運成本與系統複雜度，同時促進企業內部的敏捷開發與協作。然而，安全性與隔離性仍是推動多租戶架構時不可忽視的挑戰。本場議程將從多租戶的基本概念出發，逐步帶領大家了解實作多租戶所需考量的各個面向，並介紹如 vCluster（https://github.com/loft-sh/vcluster/）等解決方案，協助你制定符合需求的多租戶策略。
再來第二場 session，由 Shivay Lamba 帶來「Securing LLM Deployments: Isolated Model Serving with Containerization」。
隨著各界加速將大型語言模型（LLM）導入產品與工作流程，企業面臨一項全新挑戰：如何以可擴展、安全、易於維運的方式部署私有 LLM。本場議程將介紹一套雲端中立、實用的架構，說明如何透過 gVisor 與容器化實現強隔離，利用已簽章的 AI 模型確保合規與可信任的生產環境，並打造一個強化的控制平面來管理模型生命週期與基礎設施狀態。現場也將示範如何安全部署一個已簽章、容器化的 LLM，實作其在 Retrieval-Augmented Generation（RAG）流程中的導入、啟用與使用。
===
為了鼓勵大家勇於分享，2024 年 9 月以後的 CNTUG meetup 講者，我們會給予額定經費去 CNCF Store 選擇一件禮物。
制度目前試營運（可能會隨時更改），有些事項需要注意：
額定經費需要扣除當月 meetup 場地等相關費用後通知。
領取次數目前無限制，我們會優先給予沒領取過的講者分享。
禮物目前無法給予在海外的線上講者，請等待後續運作穩定後通知。
趕快按下投稿跟我們分享你的經驗吧！
https://sessionize.com/cntug-meetup/

Speakers

Hrittik Roy

Loft Labs

Platform Advocate

Shivay Lamba

Senior Software Engineer

When

Thursday, August 14, 2025
11:30 AM – 1:30 PM (UTC)

Where

五倍學院 - 實戰轉職・企業內訓．技術諮詢服務
7號 Hengyang Road
Zhongzheng District, 10046

Agenda

Multitenancy in the Kubernetes Era - By Hrittik Roy

Multitenancy is the ability to share a cluster between multiple tenants, which is an exciting concept. It involves starting with a single cluster and then allowing multiple teams to share it. This approach can be particularly useful for enterprises that want to optimize their resource utilization and reduce operational costs as well as complexity while increasing agility and collaboration. However, security and isolation can still be challenging issues that need to be addressed. This talk focuses on all the aspects that you need to know about Multitenancy before successful adoption. Starting from the basics, this talk ends on how solutions like vCluster (https://github.com/loft-sh/vcluster/) can help your multitenancy strategy.

Securing LLM Deployments: Isolated Model Serving with Containerization - By Shivay Lamba

As organizations race to integrate large language models into their products and workflows, a new requirement is emerging: the need to host private LLMs in a scalable, secure, and operationally manageable way. This talk presents a practical, cloud-agnostic architecture for scalable and secure LLM deployment. We demonstrate how to use gVisor and containerization to provide strong isolation, use signed AI models to enable compliance and production confidence and build a hardened control plane to manage LLM lifecycle and infrastructure state. The session includes a demo of deploying a signed, containerized LLM safely into this environment—showing secure onboarding, activation, and usage for a Retrieval-Augmented Generation (RAG) pipeline