---
title: "Mundher Al-Shabi, PhD (@Mundher) | Hashnode"
description: "I design and build machine learning models and systems. Read the latest articles by Mundher Al-Shabi, PhD on Hashnode."
url: "https://hashnode.com/@Mundher"
---

@Mundher

# Mundher Al-Shabi, PhD

@Mundher · Joined October 2023

I design and build machine learning models and systems.

### [Who Evaluates the Evaluator? Testing AI Simulators in Production](https://mundher.com/who-evaluates-the-evaluator-testing-ai-simulators-in-production?utm_source=hashnode&utm_medium=feed)

In my last post, I wrote about using simulators to evaluate multi-turn AI agents. It is a powerful technique, but things can get confusing fast if we do not clearly define the testing architecture. In

3 min read · [mundher.com](https://mundher.com/who-evaluates-the-evaluator-testing-ai-simulators-in-production?utm_source=hashnode&utm_medium=feed)

[1](https://hashnode.com/posts/who-evaluates-the-evaluator-testing-ai-simulators-in-production/6ad4da90529ef7522795cef0) [Oct 18](https://hashnode.com/posts/who-evaluates-the-evaluator-testing-ai-simulators-in-production/6ad4da90529ef7522795cef0)

### [Why Dynamic Reasoning Beats Static Graphs](https://mundher.com/why-dynamic-reasoning-beats-static-graphs?utm_source=hashnode&utm_medium=feed)

Standard vector RAG is great for finding needles, but it’s terrible at describing the haystack. It relies entirely on semantic proximity. That means it often fails at multi-hop reasoning or answering

3 min read · [mundher.com](https://mundher.com/why-dynamic-reasoning-beats-static-graphs?utm_source=hashnode&utm_medium=feed)

[Aug 15](https://hashnode.com/posts/why-dynamic-reasoning-beats-static-graphs/6a8032cc842b6446481159f4)

### [How to Build Privacy into LLM Agents Without Breaking Their Brains](https://mundher.com/how-to-build-privacy-into-llm-agents-without-breaking-their-brains?utm_source=hashnode&utm_medium=feed)

I've been spending a lot of time lately building autonomous agents with ReAct and RAG frameworks. The shift from stateless chatbots to stateful systems that can actually execute tools is incredibly us

3 min read · [mundher.com](https://mundher.com/how-to-build-privacy-into-llm-agents-without-breaking-their-brains?utm_source=hashnode&utm_medium=feed)

[Jul 19](https://hashnode.com/posts/how-to-build-privacy-into-llm-agents-without-breaking-their-brains/6a5cd7ccb75a471e13046e55)

### [Stacking Agent Memory: Checkpoints, Status Boards, and Active Context](https://mundher.com/stacking-agent-memory-checkpoints-status-boards-and-active-context?utm_source=hashnode&utm_medium=feed)

I love the idea of autonomous coding agents. But one of the quickest ways to hit a wall when setting them up is the infinite context problem. Your context window is finite, but the work you want the a

3 min read · [mundher.com](https://mundher.com/stacking-agent-memory-checkpoints-status-boards-and-active-context?utm_source=hashnode&utm_medium=feed)

[Jul 4](https://hashnode.com/posts/stacking-agent-memory-checkpoints-status-boards-and-active-context/6a48da4e34d1fc7d286abe50)

### [Smart LLM Routing](https://mundher.com/smart-llm-routing?utm_source=hashnode&utm_medium=feed)

Building LLM apps is easy, but scaling them without setting a pile of money on fire is hard. You really don't need the massive brainpower of GPT-5 for every single user query. Routing is how we fix th

4 min read · [mundher.com](https://mundher.com/smart-llm-routing?utm_source=hashnode&utm_medium=feed)

[Jun 26](https://hashnode.com/posts/smart-llm-routing/6a3eac60a59c1dd2c6d1e1bd)