Presented by: Oracle & MLOps Community

Oracle provides the infrastructure to build production-ready AI systems, combining structured data, vector search, and scalable database capabilities in Oracle AI Database.

Together with the MLOps Community, this workshop is focused on practical patterns for building real-world AI systems. No hype. Just technical insight, hands-on learning, and production-minded engineering.

ABOUT THE WORKSHOP

This is not your average AI workshop. It’s a hands-on session for engineers who are done building agents that forget everything the moment the session ends.

In this workshop, attendees will build a Python-based AI agent with persistent memory powered by Oracle AI Database. You’ll implement a practical memory pipeline — extract → store → retrieve → inject → forget — using structured memory and vector-based semantic recall to create agents that can retain useful context across interactions.

Along the way, you’ll learn how to encode facts, events, and preferences, design retrieval logic that surfaces the right context at the right time, and apply lightweight forgetting strategies so memory stays useful instead of noisy.

This is a practical, code-focused event built for engineers who want to move beyond demos and build AI systems that are more reliable, adaptive, and production-ready.

AGENDA

Agenda:
5:00 PM – 5:40 PM: Doors Open & Registration
Check-in, refreshments, networking, and technical setup review for the hands-on portions.

5:40 PM – 5:50 PM: Welcome & Opening Remarks
Introduction to the workshop goals, structure, and technical overview of what attendees will build.

5:50 PM – 6:55 PM: Hands-On Workshop – Part 1
Build the foundations of a memory-aware AI agent using Python and Oracle AI Database, including structured memory, storage design, and persistent memory architecture.

6:55 PM – 7:30 PM: Technical Masterclass Keynote Session
Colin Francis, LangChain, will lead a deep dive into Memory Engineering for production AI agents.

7:30 PM – 8:40 PM: Hands-On Workshop – Part 2
Implement semantic recall, retrieval logic, memory injection, and forgetting strategies to make the agent more production-ready.

8:40 PM – 9:00 PM: Wrap-Up, Q&A & Networking
Technical Q&A, networking, and next steps.

What You’ll Walk Away With

A working memory-aware AI agent
A reusable notebook and starter repo
Practical patterns for persistent agent memory
A stronger foundation for building production-ready AI systems

WHO SHOULD ATTEND

ML and AI engineers
Backend engineers building AI applications
Technical leads designing agent systems
Builders who want agents that remember, not just respond

Prerequisites: Comfortable with Python and Jupyter notebooks. OpenAI API key (or use provided alternative). Familiarity with LangChain/LlamaIndex helpful but not required.

REGISTER

Seats are limited. Register now to secure your spot.

Stop Building Goldfish Agents: Build Real Memory with Oracle AI Database