Jitendra Gupta
Enterprise Architect - Cloud & AI @ EPAM Systems
Pune, India
Actions
Jitendra Gupta is 14 years experienced IT professional with expertise in GenAI, information security and cloud computing. As a sought-after speaker, with a proven track record of implementing GenAI & security best practices in cloud environments, Jitendra is dedicated to empowering fellow IT professionals with actionable takeaways to safeguard their applications, data and infrastructure on Google Cloud.
Area of Expertise
Topics
The Agent Is the New Microservice: Rethinking Architecture for Agentic Systems
Traditional applications are built around deterministic services, APIs and predefined workflows. Agentic applications introduce something fundamentally different: components that can reason, choose tools, maintain context and dynamically decide what to do next.
So what happens to software architecture when an AI agent becomes an active participant in the system?
This session explores architectural patterns for agentic systems, including single-agent versus multi-agent designs, orchestration, tool boundaries, state and memory, event-driven workflows, human-in-the-loop controls and failure isolation.
We'll use real-world engineering scenarios to examine where agents fit naturally into existing architectures, where traditional deterministic services remain the better choice, and how to avoid turning every application into an unnecessarily complex multi-agent system.
The goal is to provide architects and experienced engineers with a practical framework for deciding when to use an agent, where to place it, and how to engineer the boundaries around it.
Spec-Driven AIOps: Building Reliable Kubernetes Intelligence with Backstage
AIOps promises automated insights and remediation for Kubernetes, but most implementations fail due to implicit assumptions, brittle automation, and lack of standardization. This talk introduces a spec-driven approach to AIOps, where operational intent is explicitly defined, validated, and executed.
We demonstrate how Backstage can act as the central control plane for AIOps by treating specifications as the source of truth for observability signals, incident workflows, and remediation actions. Using real Kubernetes scenarios, we show how specs flow from Backstage into AI-assisted analysis and deterministic automation—ensuring predictable, auditable, and safe operations.
Attendees will learn how spec-driven AIOps reduces MTTR, avoids alert fatigue, and scales across teams without embedding fragile logic into pipelines. The session includes a reference architecture, design patterns, and practical takeaways platform teams can adopt immediately.
Spec-Driven Agent Development with ADK & Antigravity
Step into the future of software engineering with Spec-Driven Development (SDD) powered by Google’s Agent Development Kit (ADK) and Antigravity. In this hands-on workshop, you’ll learn how to move beyond prompt-based coding and adopt a structured, artifact-first approach where specifications, plans, and tasks drive implementation.
Participants will build an intelligent agent by walking through the full SDD lifecycle—specify, clarify, plan, analyze, and implement—using Antigravity’s agent-first IDE. You’ll experience how AI agents collaborate with developers to generate, validate, and evolve production-ready systems with transparency and control. 
Search. Detect. Investigate. Automate: Building an Agentic SOC with Elastic, Google ADK & Google Clo
Security operations are evolving from alert-driven monitoring toward AI-assisted investigation and response. In this session, I’ll share my journey of building an Agentic SOC using Elastic, Google Cloud, and Google ADK. Through architecture insights and a live demo, we’ll see how agents correlate events, investigate threats, enrich findings, and generate evidence-backed intelligence across scenarios such as service-account compromise, privilege escalation, and data exfiltration.
Observability for AI Agents: When Your Application Can Reason — How Do You Debug It?
Traditional observability tells us what happened inside an application: requests, logs, metrics, traces and errors.
But what happens when the application itself starts making decisions?
An AI agent may call several tools, retrieve different pieces of information, change its plan, delegate work to another agent and eventually produce the wrong result — without generating a conventional application error.
This session explores observability for agentic systems, showing how to trace agent decisions, tool calls, retrieval steps, model interactions and agent handoffs. We'll examine how to answer questions such as: Why did the agent make this decision? Where did the context go wrong? Which tool call caused the failure? And how do we measure whether an agent actually performed well?
We'll also discuss practical approaches to evaluating agent quality, debugging non-deterministic behavior and designing observability into agent architectures from day one.
Leveraging Kafka on GCP for Real-Time Data Streaming in Distributed Systems
This topic will cover the best practices for deploying Kafka on GCP, focusing on scalability, reliability, and performance in distributed systems.
Low-Code IoT Integration with Google Cloud: Real-Time Data Processing and Analytics
The Internet of Things (IoT) is generating vast amounts of data that need to be processed and analyzed in real-time. This session will explore how Google Cloud’s Application Integration service can be used as a low-code solution to connect IoT devices, process real-time data, and generate actionable insights. Attendees will learn how to set up data ingestion pipelines, integrate IoT platforms, and apply real-time analytics to monitor and optimize IoT operations. A case study will illustrate the practical application of these techniques in a smart city scenario, showcasing the power of low-code solutions in the IoT ecosystem.
Integrating Neo4j with Google Gemini: Architectural Patterns for AI Agents
How can we bring the power of GenAI into graph-based applications? This talk presents proven architecture patterns for integrating Neo4j with Google Cloud’s Gemini models to build intelligent systems. We’ll show how AI agents can use Neo4j as a tool; for example, translating natural language questions into Cypher queries for graph-powered answers or leveraging a Neo4j knowledge base as contextual memory for reasoning.
Real-world scenarios will demonstrate that a graph-backed LLM handles tasks like recommendations or incident analysis more effectively than an LLM alone. We’ll also cover best practices for deploying these solutions on Google Cloud, connecting Neo4j (AuraDB or self-hosted) with Vertex AI, ensuring secure and scalable interactions, and optimizing performance.
Attendees will leave with a reference architecture for graph+AI systems and practical tips to apply these patterns—all without getting lost in theory.
Key takeaways from this session include design blueprints for graph-integrated AI agents and guidance on implementing them in enterprise cloud environments.
GitOps for Everything: Extending Declarative Principles Beyond Application Deployment
GitOps has transformed how we manage Kubernetes workloads, bringing the power of declarative infrastructure, version control, and automated reconciliation to application deployments. But why stop there? This session takes GitOps beyond its traditional boundaries to explore how declarative principles can be applied across the entire software delivery and infrastructure lifecycle.
At its core, GitOps treats Git as the single source of truth and uses automated agents to reconcile the desired and actual states of systems. While tools like Argo CD and Flux have become standard for application delivery, a new wave of GitOps patterns is extending these principles to areas such as infrastructure-as-code, cloud security policies, compliance automation, and FinOps. By leveraging GitOps for more than just app deployment, organizations gain greater consistency, transparency, auditability, and faster recovery across their tech stack.
From Spec to Action: Automating Incident Remediation with Backstage & AIOps
Incident remediation in Kubernetes environments is often slow, manual, and error-prone due to undocumented operational knowledge and inconsistent runbooks. While AIOps can detect anomalies, acting on them safely remains a challenge.
This talk presents a spec-driven approach to automated incident remediation, where operational intent is defined as structured specifications and executed through Backstage as the system of record. We show how incident signals from Kubernetes observability tools are mapped to remediation specs, enriched with AI-assisted analysis, and executed through deterministic workflows.
Using real Kubernetes failure scenarios, attendees will see how specs flow from Backstage into AIOps pipelines to trigger safe, auditable remediation actions. The session focuses on reducing MTTR while maintaining operator trust and platform safety, combining AI insights with explicit engineering controls.
From Idea to Impact: Building Agentic AI Systems the Right Way
As AI moves from reactive chatbots to autonomous agents capable of decision-making and tool usage, Agentic AI is emerging as a critical paradigm for building intelligent systems. This talk offers a deep dive into the architecture, design patterns, and implementation best practices for creating scalable and secure Agentic AI systems.
We’ll explore how to decompose complex business problems into modular, goal-driven agents using frameworks like Google’s Agent Development Kit (ADK). Through real-world use cases, you’ll learn how to orchestrate multi-agent workflows, manage agent memory and grounding, integrate external tools, and ensure alignment with enterprise standards around reliability, compliance, and observability.
Whether you're building your first AI agent or scaling an enterprise-grade agentic system, this session equips you with the strategies, frameworks, and pitfalls to avoid—ensuring your AI agents are not only autonomous but accountable and aligned with business goals.
From Idea to Impact: Building Agentic AI Systems the Right Way
As AI moves from reactive chatbots to autonomous agents capable of decision-making and tool usage, Agentic AI is emerging as a critical paradigm for building intelligent systems. This talk offers a deep dive into the architecture, design patterns, and implementation best practices for creating scalable and secure Agentic AI systems.
We’ll explore how to decompose complex business problems into modular, goal-driven agents using frameworks like Google’s Agent Development Kit (ADK). Through real-world use cases, you’ll learn how to orchestrate multi-agent workflows, manage agent memory and grounding, integrate external tools, and ensure alignment with enterprise standards around reliability, compliance, and observability.
Whether you're building your first AI agent or scaling an enterprise-grade agentic system, this session equips you with the strategies, frameworks, and pitfalls to avoid—ensuring your AI agents are not only autonomous but accountable and aligned with business goals.
Containers Are Disposable—Your Data Isn't: Stateful Workloads Done Right
Containers are designed to be ephemeral, but the applications they power often rely on persistent, business-critical data. This session explores how to successfully run stateful workloads in Kubernetes without compromising reliability, performance, or data integrity. Through real-world architectures and live demonstrations, attendees will learn how Kubernetes manages persistent storage using Persistent Volumes, Persistent Volume Claims, StorageClasses, and CSI drivers, along with best practices for deploying databases, message brokers, and AI vector stores. The talk also covers backup and disaster recovery with Velero, high availability patterns, storage performance optimization, and production considerations on cloud platforms. Whether you're modernizing legacy applications or building cloud-native AI platforms, this session will equip you with practical strategies to ensure that while containers may come and go, your data remains secure, resilient, and always available.
Build a Mini AI Engineering Team with Agents
What if instead of asking one AI to build your application, you could give different AI agents different jobs — just like a real software engineering team?
In this session, we'll build a mini AI engineering team where specialized agents collaborate to solve a software development task. A Planner Agent breaks down the problem, a Developer Agent works on the implementation, a Tester Agent validates the solution, and a Reviewer Agent checks the result.
We'll explore how these agents communicate, share context, use tools and pass work between one another. More importantly, we'll examine whether adding more agents actually makes a system better — and the challenges that come with multi-agent architectures, including coordination, context management, reliability and cost.
The session will provide a practical introduction to multi-agent systems through a developer-focused example, helping attendees understand when a single agent is enough and when multiple specialized agents can make sense.
No advanced AI/ML knowledge is required.
AgyQueue: Building Non-Blocking Multi-Agent AI Systems with Durable Workflows
In this session, we’ll introduce AgyQueue, an open-source asynchronous execution framework that enables AI agents to execute long-running workflows without blocking users. Built for cloud-native environments, AgyQueue combines durable workflows, distributed task queues, checkpoint recovery, and real-time progress tracking to orchestrate autonomous multi-agent systems.
MCP: The USB-C of AI — Connecting AI Agents to the Real World
An LLM can write code, explain concepts and answer questions — but what happens when you want it to use your tools and actually get things done?
This session introduces Model Context Protocol (MCP) through a simple developer-first approach. We’ll explore how MCP enables AI applications and agents to connect with external tools, data and services through a common interface — much like USB-C made it easier to connect different devices.
Starting with a simple AI application, we’ll progressively give it access to real-world capabilities such as files, APIs, developer tools and application data. We’ll see what happens when an AI moves from simply generating an answer to using tools, gathering information, reasoning over the results and taking action.
The session will also cover the basic architecture of MCP, how MCP servers and tools work, and where this technology fits into the rapidly evolving agentic AI ecosystem.
No deep AI or ML background is required. If you can build a basic application and are curious about what comes next for AI, this session is for you.
Unlocking Cost Transparency - Google Cloud's Cost Attribution Solution for Effective FinOps
This session will introduce attendees to Google Cloud's Cost Attribution Solution, focusing on how organizations can achieve granular visibility into their cloud spend and optimize resources using FinOps best practices.
We’ll demonstrate the importance of labeling strategies and governance, using Google Cloud’s built-in capabilities, along with proactive and reactive governance approaches for cost management.
Unlocking Cost Transparency - Google Cloud's Cost Attribution Solution for Effective FinOps
This session will introduce attendees to Google Cloud's Cost Attribution Solution, focusing on how organizations can achieve granular visibility into their cloud spend and optimize resources using FinOps best practices.
We’ll demonstrate the importance of labeling strategies and governance, using Google Cloud’s built-in capabilities, along with proactive and reactive governance approaches for cost management.
Security Monitoring on Google Cloud
Learn how to effectively monitor the security of your applications and infrastructure on Google Cloud. I will cover best practices for setting up logging and monitoring, leveraging tools like Cloud Logging, Cloud Monitoring, and Cloud Security Command Center to detect and respond to security incidents.
From understanding how to configure security alerts to analyzing logs and implementing proactive security measures, this talk will provide insights and strategies for robust security monitoring on Google Cloud, helping you safeguard your applications and data from potential threats.
Google Cloud Security Advancement with Generative AI
In the ever-evolving realm of technology, the fusion of Google Cloud Security and Generative AI promises a future where our digital assets are safeguarded more efficiently than ever before. This talk will dive deep into the intricate synergy between Google Cloud's robust security features and the limitless potential of Generative AI. We will explore real-world scenarios and the transformative impact this union can have on businesses. Understand the potential threats of tomorrow and discover how these advanced technologies are shaping a safer digital world for us all.
Google Cloud Landing Zone - Design Best Practices & Demo
Landing Zone is very important part of any cloud migration or build from stretch projects.
Almost a year back when I looked for relevant resources for GCP Landing Zone, I could hardly find anything apart from Google Cloud documentations.
I filled this gap by producing my own YouTube videos/medium articles to explain GCP Landing Zone, also presented it at Google Cloud Community Day-2023, Bhopal.
As part of this talk I am going to add a Demo of Landing Zone creation using Terraform.
Google Cloud FinOps Best Practices for Generative AI
Navigating the landscape of cloud finances can be complex, especially when integrating advanced technologies like Generative AI. This talk delves into best practices for managing costs, optimising resources, and ensuring maximum ROI when employing Generative AI on the Google Cloud platform. I will discuss the nuances of FinOps (Financial Operations) in the cloud environment, share insights into effective budgeting and cost management, and provide actionable strategies. Equip yourself with the knowledge to harness the power of Generative AI without breaking the bank. Discover how to strike the perfect balance between innovation and fiscal responsibility in the cloud era.Google Cloud FinOps best practices for Generative AI
Generative AI for Automated Kubernetes Policy Management
As Kubernetes environments scale, managing policies becomes increasingly complex and time-consuming. This talk delves into the cutting-edge application of generative AI for automating the creation, optimization, and enforcement of Kubernetes policies.
By leveraging AI-driven insights, attendees will discover how to predict and mitigate security risks, ensure compliance, and optimize cluster configurations with minimal manual intervention.
This session will provide practical examples and real-world case studies to illustrate how generative AI can enhance operational efficiency and security in cloud-native environments, offering a glimpse into the future of autonomous Kubernetes management
Grounding generative AI in enterprise truth for real-world impact
In general, the most common challenges that come up relate to data quality and management. Like any data or AI system, the success of RAG relies heavily on how well your data is organized and maintained, and how well you understand what it is and how it relates to your overarching objectives and goals.
In particular, many companies overlook the importance of data hygiene when they start implementing RAG components. The old computer science adage of “garbage in, garbage out” still holds true in the gen AI era; if your data is incomplete, inconsistent, or inaccurate, the information RAG retrieves will be too. For instance, if you have two pieces of data in your dataset that contain varied information for the same query, you might end up getting a different answer every time that question is asked.
Designing and Orchestrating Multi-Agent AI Systems Using Google’s Agent Development Kit (ADK)
This session demonstrates how to design and orchestrate multi-agent AI systems using Google’s Agent Development Kit (ADK). It showcases how complex tasks can be decomposed across specialized agents coordinated through sequential, loop, and parallel workflows. Participants learn to manage agent hierarchies, shared state, and execution flow for reliable outcomes. The approach highlights how ADK enables scalable, modular, and production-ready agentic applications on Google Cloud.
Designing Agentic Terraform Workflows with Google ADK
Infrastructure automation is evolving beyond scripts and pipelines. This talk introduces agentic Terraform workflows using Google’s Agent Development Kit (ADK). We’ll explore how multi-agent patterns like orchestrators, parallel agents, critics, and human-in-the-loop approvals can generate, validate, and govern Terraform at scale. Attendees will see how agent teams replace brittle pipelines with adaptive, auditable infrastructure workflows.
DevFest Chandigarh 2024 Sessionize Event
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top