תוכנית Oracle AI Week 2026

The Agent Stack Behind Reliable Agentic Systems

Agents fail in production for reasons that rarely show up in a demo: lost context, duplicated work, runs that die halfway and start over.
The model isn't the problem. The harness is. The agent harness is the engineering layer wrapped around the model: it plans, fans work out, verifies results, checkpoints.
This session is a working tour of the harness's memory surfaces, the places where memory engineering decides whether an agentic application can be trusted with real work. Reliability lives in this layer, and every one of those capabilities resolves to a memory operation. progress, and decides what enters the context window.

Modern AI Agent Memory: Methods, Algorithms, and Challenges

In this session We will discuss the main challenges that agent memory is designed to address, the algorithmic layer used to improve memory efficiency and coherency, as well asexamples from several leading scientific publications.
We will also explore the key challenges currently facing this field.

Observability in the AI Era: From Telemetry to Trusted Action

In this session, we will explore how AI systems emit richer evidence and how AI transforms raw telemetry into operational intelligence.

Foundations: LLMs, Agents & MCP

A comprehensive overview of LLM foundations, agent loops, and tool-use architecture.
Learn why agents need external capabilities and how Model Context Protocol (MCP) standardizes agent-environment interactions.

Using Existing MCP Servers

A practical, demo-driven session focusing on connecting clients to existing MCP servers.
Through live demonstrations, participants will learn how to configure environments, discover tools and resources, execute multi-tool calls, and troubleshoot real-world connection issues.

Building Your Own MCP Server

A step-by-step live demo on building a custom MCP server in Python.
Participants will observe project scaffolding, custom tool and resource implementation, and client integration with error handling.

Infrastructure as Code: Automation Without Losing Grip

“Create another environment exactly like production.” Which version of production?
An urgent console fix, a copied configuration and a script that stops halfway can quickly turn infrastructure into a guessing game. Automation helps—but how do we make changes repeatable without losing visibility, ownership and control?
This session follows the journey from manual work → scripts → Terraform → Terragrunt → a full GitOps ecosystem. Through practical examples and guided walkthroughs, we’ll explore reusable infrastructure, environment management, drift detection and controlled execution with OCI Resource Manager. We’ll then connect infrastructure provisioning with Helm and Argo CD, examining how GitOps changes deployment, reconciliation and recovery.
We’ll also discuss Karpenter for OCI and Crossplane, focusing on the ownership boundaries needed when multiple controllers manage infrastructure.

Audience: DevOps and platform engineers, SREs, cloud architects and developers responsible for infrastructure. Familiarity with cloud fundamentals is recommended; Terraform and Kubernetes knowledge is helpful.

Failure Occurs Only in Production

The servers are green. The customer is still waiting. What do you investigate first?
Production brings together traffic, dependencies, configuration and timing in ways that testing rarely reproduces. When something fails, more dashboards do not automatically lead to a better explanation.
Using an illustrative production incident, this session connects metrics, logs and distributed traces to investigate competing explanations, identify bottlenecks and verify recovery.
We’ll examine how an apparently slow database operation can actually be time spent waiting for a shared resource - and what evidence separates those possibilities.
Product walkthroughs and screenshots show where OCI Application Performance Monitoring (APM), Oracle Log Analytics and OCI Monitoring fit, alongside OpenTelemetry, Prometheus, Grafana, Loki, Tempo and Jaeger. We’ll also translate user expectations into service indicators, service targets and error budgets.

Audience: Developers, DevOps engineers, SREs, DBAs and architects involved in diagnosing production systems. Basic familiarity with APIs and cloud applications is recommended.

Resilience by Design: Building Resilient Applications on OCI

When critical applications go down, the impact goes far beyond IT. Resilient platforms keep critical services running, protect data, and enable rapid recovery when disruption occurs.
This seminar explores how resilience is designed across the OCI stack, covering high availability, networking, data protection, RPO/RTO, and cross-region disaster recovery, followed by a live OCI Full Stack DR demonstration.

Audience: Cloud and infrastructure architects, application architects, database administrators, DevOps and platform engineers, and IT professionals responsible for service availability, business continuity, and disaster recovery.
No mandatory prerequisites; basic familiarity with cloud concepts is required.

Single-Agent Bottlenecks

Focus: Understanding single-agent mechanics, tool overload, and basic state graph architecture.

Scaling with Multi-Agent Patterns and Middlewares

Focus: Sub-agent orchestration, handoffs, skills routing, and state interceptors.

Advanced Multi-Agent Engineering

Focus: Deep LangGraph workflows, full architectural refactoring, and cost/context optimization.

From Writing to Judging: Software Engineering in the Age of AI Code Execution

Topics Covered:
• The New Bottleneck.
• The "Confidence Trap".
• Critical Code Review for Machine Output.
• Practical Failure Modes.

Concurrency, Algorithms & Architecture Where AI Fails (and You Need to Rescue It)

Topics Covered:
• Tests as the Leash.
• Behavior vs. Implementation.
• Evolving Agile Processes.
• Modern AI Workflow Lifecycle.

The AI Leash Building an Enterprise Safety Net and Rethinking Development Processes

Topics Covered:
Tests as the Leash.
Behavior vs. Implementation.
Evolving Agile Processes.
Modern AI Workflow Lifecycle.

Building a Production – Grade AI Virtual Assistant

How do you evolve from a simple LLM-based chatbot into a scalable AI platform combining conversation, knowledge, speech, and real-time interaction?
Based on three years of hands-on experience designing and running the AI backend for Sidekick Platform, this session presents the architecture behind a modern Virtual Assistant and the key decisions required to operate it at scale.

Operating a Production AI Platform – Models, Prompts, Configuration and Lessons

Production AI systems quickly move beyond a single model and a few hardcoded prompts. Models change, providers evolve, prompts become business-critical assets, and the backend itself requires strong operational control, configuration, observability and administration.
This session presents practical architecture and best practices for managing models, prompts and the operational backend of a production AI platform, based on lessons learned from real-world use.

RAG in the Real World – What We Learned After Putting It into Production

RAG often looks simple in a diagram: create embeddings, search a vector database, and send the results to an LLM. In production, retrieval quality quickly becomes one of the hardest and most important parts of the AI architecture.
This session follows the practical evolution of RAG based on real production experience.

Oracle Integration Foundation for Enterprise AI Agents

How do we connect AI agents to enterprise systems and enable them to perform business actions under control?
This lecture will present Oracle Integration as the infrastructure for connecting applications, data and AI services. It will review MCP, AI Assistant and Human-in-the-Loop for creating intelligent business processes with human oversight.
Key topics include:
Connecting systems and exposing integrations as tools for AI agents via MCP
Managing access to MCP servers through MCP Gateway
Assisting in the development and documentation of integrations using AI Assistant
Incorporating human approvals and handling exceptions via Human-in-the-Loop
Demonstrating a business process that combines an AI agent, integration, and human control.

This lecture is designed for architects, integration and application developers, and technology leads with basic familiarity with APIs and system-to-system integration. No prior experience in AI agent development is required.

Participants will gain understanding of the role of Oracle Integration in connecting AI agents to existing systems and incorporating human control into business processes.etails soon...

Agentic Development with Oracle Database, APEX & MCP

Practical AI with Oracle Database, SQLcl MCP and APEX

AI-assisted development is evolving rapidly, while Oracle Database and APEX provide the stability, security, and structure required for enterprise applications.
This session explores how Oracle development teams can combine the best of both world-using AI agents to accelerate and enhance development while continuing to rely on the strengths of the Oracle Database and APEX’s declarative, low-code model.

We’ll demonstrate practical ways to integrate AI agents throughout the development lifecycle, including generating, reviewing, and testing SQL and PL/SQL, interacting with the database through SQLcl MCP, and leveraging new Oracle APEX capabilities such as application creation with APEXlang and built-in AI agents.
We’ll also show how AI can help build tailored APEX solutions using PL/SQL, HTML, CSS, and JavaScript, create rich visual and interactive components, and extend the platform beyond standard declarative features.

The focus is on practical, controlled adoption of AI-selecting the right use cases, providing the right business and technical context, integrating with Git and CI/CD workflows, and validating generated code-to accelerate development, modernization, and prototyping without compromising security, quality, maintainability, or developer control.

Oracle Data Base AI: The Big Picture

In this session we will cover major components of AI capabilities in Oracle DB.
• MCP Server
• Vectors & RAG, Vector Indexing
• Enterprise Scale AI
• AI Profile
• Comments and Annotations
• SelectAI (NL2SQL, Synthetic data, Text Summarization, Translations, External Data Gateway, AI Agents, Multi Agents Workflows…)
• Private Agent Factory
In this session we will provide an introduction. Some of the topics will be covered in deep dive sessions.

Select AI in Oracle DB

In this session we will cover major components of AI capabilities in Oracle DB.
• MCP Server
• Vectors & RAG, Vector Indexing
• Enterprise Scale AI
• AI Profile
• Comments and Annotations
• SelectAI (NL2SQL, Synthetic data, Text Summarization, Translations, External Data Gateway, AI Agents, Multi Agents Workflows…)
• Private Agent Factory
In this session we will provide an introduction. Some of the topics will be covered in deep dive sessions.

Private Agent Factory (PAF)

Private Agent Factory is a free component that can be set both as part of OCI Marketplace and as image on your local or cloud environment.
It requires access to Oracle DB and some LLM and helps user to easily create and provide access to AI agents.
It comes with easy-to-use UI for AI Agents creation, maintenance and a built-in runtime UI.
Many thousands of production Agents are running worldwide based on Private Agent Factory. In this session we will explain the basic setup of PAF and create (and run) various Agents with it.

Oracle Integration Foundation for Enterprise AI Agents

Foundation: How The Engine Executes Your Query

This session takes you from SQL text to a plan tree you can read and cost, ending on the join algorithms behind most of it.

Performance Optimization Techniques

This session is about the levers: what an index really costs, how to design one the planner will use, when partitioning pays, and the estimates that decide all of it.

AI In Performance Optimization

This session is the AI part: why a model reading only your query text guesses, what evidence turns it into an analyst, the read-only MCP tool surface, the agentic tuning loop and what an agent may and may not do to production.

Technology Trends & Roadmaps

The strategic view: which technology shifts are real, which are marketing and how to turn that into a funded plan.
You will leave with a structured roadmap template and a clear view of where to spend and where to wait.

The New Offensive Landscape

Attackers did not get smarter. They got cheaper, faster and more numerous - and the AI tools your organization already deployed have themselves become targets.
You will leave with a clear picture of which threats have genuinely changed and which have not and a set of questions to put to your own team.

Defense, Governance & Resilience

Defenders cannot respond at human speed alone - but "add AI to the security team" is a slogan, not a strategy.
You will leave with a practical checklist and a view of which obligations apply to you.

The Agentic Enterprise: Reimagining Business Processes with AI

This session explores the shift from Generative AI to Agentic AI, demonstrating how autonomous agents execute end-to-end business actions directly within enterprise applications.
Designed for both business leaders (process managers, operations, innovation leads) and technology professionals (architects, developers, AI leaders).
No prior AI agent development experience or personal laptop required; general familiarity with business processes is helpful.

From Prompts to Action: Advanced Generative AI for Project and Technology Leader.

Generative AI can do much more than answer questions or create content.
This practical session shows how to turn AI into an active work partner that can organize information, plan work, create repeatable processes, check its output and prepare recurring deliverables.
We will use one end-to-end example while keeping human judgment and approval at the center.
Designed for project managers, technology leaders, consultants and service delivery teams managing complex workflows, no coding experience is required.
Participants only need basic familiarity with GenAI tools.

Oracle Analytics & AI, what works and how

In this session, we will have a very short overview of Oracle Analytics, Both Analytics Server (OAS) and Cloud (OAC).
Most of the session will cover the AI oriented capabilities in the product. Most of them are currently available only on OAC. Some on OAS as well.
If you want a short overview of Oracle Analytics and then have a deep dive into Chatbots generation, AI Assistant, MCP access to Analytics, AI generated formulas and many more… This is the session for you.