Tal Waitzenberg

Talks & Lectures

I speak about building AI systems for security: autonomous investigation, agent architectures, and evaluation that reflects real-world behavior. My talks draw on practical work in threat detection, cloud security, and production AI.

Topics: security agents · A/A and shadow testing · context management · ML for threat detection

Selected talks

Same Threat, Two Verdicts: Shadow Testing Our SOC Agent

[un]prompted II · San Francisco · 2026
With Erez Harush. Measuring agent variability, separating evidence from judgment, and evaluating changes before they reach production. Conference agenda.

Deep Threat Investigation Using Autonomous AI Agents

LangTalks AI Engineering Conference
Context management, planning, and evaluation for agents that investigate security threats. Session details.

Tal Waitzenberg presenting context isolation and sub-agents at LangTalks
On stage at LangTalks, discussing context isolation and sub-agents.

Wiz AI Agents: Mika AI & Wiz MCP

Google Cloud
Connecting AI agents with security context and tools to support cloud security workflows.

Tal Waitzenberg speaking about Mika AI and Wiz MCP at a Google Cloud event
Presenting Mika AI and Wiz MCP at a Google Cloud event.

LLM evaluation in practice

AWS GenAI Loft · Tel Aviv
Using LLMs to evaluate conversations and responses, and turning evaluation into a practical feedback loop.

Tal Waitzenberg presenting LLM evaluation at AWS GenAI Loft in Tel Aviv
Discussing LLM-as-a-judge evaluation at AWS GenAI Loft.

Earlier lectures

Tal Waitzenberg speaking at a meetup

LangTalks 1st Meetup lecture at Microsoft Reactor, Tel-Aviv.

Tal Waitzenberg giving a lecture

The Road From AI to AGI lecture at Cybereason, Tel-Aviv.

Lecture at Unity

Using ChatGPT in Day-to-Day Work lecture at Unity, Tel-Aviv.

Tal Waitzenberg presenting to an audience

For speaking invitations: talon8080@gmail.com.