AI Lab (SF) - Build Reliable Agents with LangChain and Oracle AI Database
First 160 characters summarize the event for meta purposes: Learn to build reliable AI agents with LangChain and Oracle AI Database in a hands-on SF workshop by Oracle and AICamp. Overview Oracle and AICamp present a focused, interactive workshop in San Francisco designed for developers who want to elevate AI agent reliability beyond what a single-turn interaction can offer. The session targets memory-centric agent design, showing how to design, implement, and evaluate memory-enabled agents that maintain continuity, reduce drift, and stay grounded over multiple turns. This is a practical, end-to-end lab: from memory concepts to a working reference implementation you can reuse in your own projects. What you’ll learn and do - Understand Agent Memory and why it matters beyond ordinary chat history - Build practical memory abstractions: short-term working memory, semantic memory, episodic memory, and lightweight procedural memory - Implement ingestion, storage, retrieval, summarization, and managed forgetting patterns to keep memory fresh and useful - Integrate real-time internet lookups via Tavily and use Oracle AI Database as the agent’s Memory Core for unified retrieval and persistence - Apply context engineering, prompt engineering, and memory engineering to reduce drift, improve grounding, and maintain consistency across turns - End with a working reference implementation and reusable patterns for scalable agent scaffolding and evaluation harnesses focused on reliability and believability Who should attend This hands-on workshop is ideal for AI developers, AI engineers, AI practitioners, and Python developers who want to design more robust AI agents and explore memory-centric architectures. Format and logistics - Date and time: Thu, Feb 19, 2026, 5:30 PM (local time, Pacific) - Location: San Francisco, CA (in-person) - Registration is required on an external site for admission - Organisers: Oracle and AICamp in collaboration with the SF Bay AI community; the meetup page notes a large, active community (SF Bay Area and global reach) - Registration/Topic call for speakers: topics welcome via the event page Why this stands out If you’ve built AI agents that perform well in a single-turn chat but falter across sessions, this workshop provides a concrete framework for turning stateless loops into memory-augmented agents. Expect deep dives into memory architectures, practical design patterns, and a reproducible setup you can adapt to your own projects. Ex

