The Fifth Elephant - Pune Edition

AI Agents can autonomously think, plan and execute with minimal human supervision. AI Agents have shown promising results in improving productivity which is why enterprises are eager to adopt AI agents. AI Agents are designed to handle complex tasks using the underlying model’s reasoning capabilities which are context sensitive and probabilistic. This means sometimes, they can exhibit non-deterministic behaviours.

Enterprises can adopt AI Agents only after addressing the Security, Compliance and Operational concerns with them. They would need full visibility and control over what AI Agents do. There have been cases where Agents have leaked sensitive information or misconfigured environments. Ensuring AI Agents comply with the organization’s IT and Data Access policies is necessary. Building an AI agent that works in a demo is easy. Building one that enterprises can trust in production is a different challenge altogether.

In this workshop, we’ll start with the fundamentals of AI agents, LLMs as the reasoning engine and tools as the execution layer. We will then focus on what separates an enterprise-ready deployment from a proof-of-concept: authentication, role-based access control (RBAC), rate limiting, observability and audit trails. And how these will help enterprises overcome the challenges with AI Agent Adoption. You’ll learn how to leverage Nutanix Enterprise AI (NAI) to deploy production-grade AI agents with minimal operational overhead. We’ll walk through the platform’s streamlined workflows for Model Deployment and Access Control, Secure Inference API Access, MCP (Model Context Protocol) Server Integration, and Regulated MCP Tool Permissions.

By the end of this workshop, you’ll have a clear blueprint for deploying AI agents that meet enterprise security and compliance requirements, maintaining developer velocity.

Key takeaways

• Understand the core principles of AI agents and how they function.
• Attendees will leave with hands-on experience of writing a MCP Server, writing a MCP Client along with a Agent Loop. We will build a Code Review Agent using the MCP Server and gpt-oss-120b model.
• Along with the hands-on knowledge, attendees will be exposed to the reasons why Enterprises despite being eager are skeptical about adopting AI Agents.
• Attendees will leave with a deeper understanding building Enterprise Grade MCP Servers addressing the problems discussed.

Level: Beginner/Intermediate

Prerequisites:

• Basic understanding of LLMs and how Inference works
• Basic understanding of MCP
• Hardware/Software requirements:
  • Mac or Windows computer
  • Latest version of Python (preferably 3.12.6) and uv installed

Target audience

• Software engineers building AI applications for enterprise use
• Engineering managers and architects designing secure AI Agents and AI Platforms
• IT Admins managing AI Infrastructure and MCP Servers

Workshop Plan

Outline

1. Introduction to Agents, Inference, AI Infrastructure Complexities (10 minutes)
2. Model Deployment using Nutanix Enterprise AI (20 minutes)
3. Writing a Traditional AI Application (30 minutes)
4. How to enrich the context of an LLM? What is MCP? What are the challenges in deploying AI Agents? (10 minutes)
5. Writing a MCP Server (30 minutes)
6. Securely Connecting to a MCP Server using Nutanix Enterprise AI (20 minutes)
7. Traditional AI Application + Tool Capabilities = AI Agent (1 hour)

Workshop Setup

Github Repository: AgentOps

Instructors

Aishwarya Raimule is a ML Systems Engineer at Nutanix, India where she is currently working on Observability for Nutanix AI Inference Platform for LLMs and MCP Integration for Agentic workflows.

Hritik Raj is a ML Systems Engineer at Nutanix since a year and half. He has been working on AI observability and LLM benchmarking initiatives. He is interested in Agentic orchestrators, Finetuning and LLM evaluations.