Kubernetes Bangalore Meetup

KServe is a mature platform for model serving. However, securing machine learning inference remains a fundamental challenge. ML models are like a black box. One can not just easily find if a model has been poisoned or tampered with.

The talk begins with real-world case studies of ML security failures highlighting common attack vectors and production grade security issues in ML systems.

Next, we introduce KitOps, a CNCF project that enables a zero-trust approach to ML inference on KServe. By using ModelKits: signed, OCI-native bundles that package models, metadata, configurations, and dependencies, we can establish verifiable trust throughout the model lifecycle. ModelKits can be stored in any OCI-compatible registry and seamlessly integrated into existing CI/CD pipelines.

Finally, we will walk through production patterns for hardening KServe deployments using familiar cloud-native tools: cryptographic model signatures, policy engines for enforcement, and AI Bill of Materials (AIBOM) for auditability and compliance.

Thus you will learn how to harden KServe using the similar set of cloud native tools already at your disposal.

Debugging Kubernetes issues require a bit of expertise with understanding Kubernetes and what all data is accessible where. For an issue, you might need to look infrastructure metrics, application metrics, application logs, kubernetes events, traces etc. Oodle AI is a platform which ingests all this data and provides both easy navigation across these signals for a human as well as ability to troubleshoot Kubernetes issues via AI agents faster and more reliably. I will present a demo of few common Kubernetes troubleshooting and how easy it is to debug them with Oodle AI.

As Kubernetes platforms scale, manual reviews and tribal knowledge no longer work. Teams need automated guardrails that enforce security, reliability, and compliance—without blocking developer velocity. In this talk, we’ll explore how to build a policy-driven Kubernetes platform using Kubernetes-native controls. You’ll learn how platform teams can: • Enforce environment isolation and security baselines • Prevent misconfigurations before they reach production • Standardize workloads across teams and clusters • Shift governance left into the deployment lifecycle

The session focuses on practical patterns, real-world lessons, and common mistakes observed in production Kubernetes environments.

This talk is ideal for teams moving from “Kubernetes as infrastructure” to Kubernetes as a product.

The talk explores how data flows through an ML/AI system—from ingestion and preprocessing to scalable storage and retrieval. We’ll cover design patterns, batch vs streaming pipelines, feature stores, vector databases, and data lakes, with a focus on building reliable, cost-efficient, and production-ready AI platforms.

Why rely on expensive cloud credits or centralized clusters? In this talk, we’ll demonstrate how to create a fully functional, cloud-like environment using only open-source tools, right on your laptop. We use k3d / kind, ArgoCD, OpenTelemetry, Grafana, NGINX ingress, and Keycloak to simulate a modern production-grade cluster.

This talk walks through cluster setup, GitOps deployment, monitoring, service mesh options, autoscaling, and debugging: all from a developer machine. Perfect for students, indie hackers, and engineers who want to experiment with cloud-native architectures without paying for AWS/GCP.

Key Takeaways How to set up a Kubernetes cluster locally in minutes. Automating deployments with ArgoCD (GitOps). Adding monitoring, tracing, and ingress the FOSS way. Local-first workflows that mirror production environments. How to practice DevOps/SRE skills with zero cloud cost.