

RedisTLV: Building Stateful AI Systems
โAll talks will be delivered in Hebrew.
โโ17:30-18:00 Pizza, beers and networking ๐๐บ
โ18:00-18:30 From the Open Web to Redis: Context for Stateful Agents
โStateful agents need two things - memory of what they've seen, and access to what's happening right now. Most teams build the memory and skip the web. I'll show how Bright Data turns the open web into clean, structured signal that agents can actually use, and how that signal pairs naturally with Redis for caching, vector search, and conversation memory. Concrete patterns, real numbers from a production MCP server, and a clear take on where each piece of agent context belongs.
โSpeaker: Meir Kadosh, AI Engineer at BrightData
โ18:30-19:00 Building Agents That Actually Remember
โWe design AI agents with human-like personalities and emotions, yet they often suffer from "memory lapses," reverting to the limitations of their last token. In this session, weโll move beyond the "persona" to explore how to build persistent, institutional memory for our agents.
Weโll bridge the gap between human-like behavior and technical reality by covering:
- The Psychology of Memory: Why AI struggles to maintain context compared to human episodic memory.
- Evolving Memory Layers: How to use Cognee to ingest, structure, and reason over unstructured data.
- Production-Ready Memory: Using Redis as a stateful memory server to implement custom forgetting policies and shared context across agents.
โSpeaker: Amit Rosen, AI Solutions Architect at Keshet Media Group
โ19:00-19:30 RAG is not enough; Introducing Context Surfaces
โWeโll introduce Redis Context Engine - a new way to expose structured, real-time data to AI agents. Including Retriever, Memory and Caching for agents.
Weโll cover:
โWhy unstructured context breaks down in production
โHow agents move from retrieval โ tool-driven reasoning
โTurning data models into queryable interfaces (MCP)
โWhat this means for building reliable AI systems
โSpeaker: Itay Tevel, Principal Solutions Architect at Redis