AI Dev 26 X SF | Eli Schilling: Hands On Agent Context & Memory Engineering with Oracle AI Database

Why It Matters

Unified memory in Oracle’s AI Database lets enterprises build scalable, context‑aware agents that cut token costs and deliver consistent, long‑term assistance across teams.

Key Takeaways

• Oracle AI Database supports unified multi‑modal memory for agents.
• Short‑term, long‑term, procedural memory reduce token bloat and hallucinations.
• Open‑source code and Docker images enable rapid agent memory prototyping.
• Memory layers enable consistent enterprise‑wide agent behavior across teams.
• Upcoming 4‑hour deep‑dive workshop offers hands‑on agent memory engineering.

Summary

Eli Schilling’s talk at AI Dev 26 focused on building robust memory architectures for autonomous agents using Oracle’s AI Database. He outlined how a unified, multi‑modal database can store relational, vector, graph, and spatial data, eliminating the need for disparate storage systems.

The core insight is that agents require layered memory—short‑term session context, semantic caches, and long‑term procedural memory—to avoid token bloat, hallucinations, and repetitive prompts. By persisting knowledge across days, weeks, or months, agents can reason more like humans, leveraging tools and external resources without re‑querying large contexts each time.

Schilling emphasized, “An AI agent without memory is just autocomplete with ambition,” and illustrated how procedural memory mirrors a human expert’s instant recall of recurring solutions. He demonstrated a code‑first approach, offering a GitHub repo, Docker container, and Oracle DB setup for developers to experiment immediately.

The significance lies in enabling enterprises to deploy consistent, scalable agents that learn and adapt over time while simplifying infrastructure overhead. The free deep‑learning course and upcoming four‑hour workshop provide hands‑on training, accelerating adoption of memory‑driven AI workflows.

Original Description

At AI Dev 26 x San Francisco, Eli Schilling provided a technical overview of architecting memory and context systems for autonomous AI agents, emphasizing the use of Oracle Database as a unified platform for managing various data modalities.