---
title: "Amit Sides (@amitsides) | Hashnode"
description: "AI Engineer. Read the latest articles by Amit Sides on Hashnode."
url: "https://hashnode.com/@amitsides"
---

@amitsides

# Amit Sides

@amitsides · Tel Aviv, Israel · Joined September 2024

AI Engineer

[Website](https://www.linkedin.com/in/amit-sides-85251445/)

### [A Multidisciplinary Theory of Exploiting LLMs and Guardrails: The Collapse of Embedding Geometry](https://ai-blog.dev/a-multidisciplinary-theory-of-exploiting-llms-and-guardrails-the-collapse-of-embedding-geometry?utm_source=hashnode&utm_medium=feed)

Multidisciplinary red team research on LLM SecOps reveals new vector attacks: Adversarial Vector Semiotics, Epistemologic Poisoning, CryptoLinguistics

13 min read · [ai-blog.dev](https://ai-blog.dev/a-multidisciplinary-theory-of-exploiting-llms-and-guardrails-the-collapse-of-embedding-geometry?utm_source=hashnode&utm_medium=feed)

[Sep 21, 2025](https://hashnode.com/posts/a-multidisciplinary-theory-of-exploiting-llms-and-guardrails-the-collapse-of-embedding-geometry/68d027855a74ece3f115183b)

### [Agentic AI SecOps Overview Part 1: Between Red-Blue Multi-Agents ContextSecOps Bottle and Poison](https://ai-blog.dev/agentic-ai-secops-overview-part-1-between-red-blue-multi-agents-contextsecops-bottle-and-poison?utm_source=hashnode&utm_medium=feed)

This article overviews Agentic AI SecOps problems and challenges for 2026: Context Guardrails

11 min read · [ai-blog.dev](https://ai-blog.dev/agentic-ai-secops-overview-part-1-between-red-blue-multi-agents-contextsecops-bottle-and-poison?utm_source=hashnode&utm_medium=feed)

[Aug 25, 2025](https://hashnode.com/posts/agentic-ai-secops-overview-part-1-between-red-blue-multi-agents-contextsecops-bottle-and-poison/68abe8b7a1f67613fcfacc7c)

### [AI Agents and their Ontologies, Psychologies, RAG Capabilities, Identity, Personality, Communication Skills, and Infrastructure](https://ai-blog.dev/ai-agents-and-their-ontologies-psychologies-rag-capabilities-identity-personality-communication-skills-and-infrastructure?utm_source=hashnode&utm_medium=feed)

Orchestrating Micro Agents to communicate with each other with underlying infrastructure: SQL, Graphs & Vectors

10 min read · [ai-blog.dev](https://ai-blog.dev/ai-agents-and-their-ontologies-psychologies-rag-capabilities-identity-personality-communication-skills-and-infrastructure?utm_source=hashnode&utm_medium=feed)

[Mar 29, 2025](https://hashnode.com/posts/ai-agents-and-their-ontologies-psychologies-rag-capabilities-identity-personality-communication-skills-and-infrastructure/67e7f18fc8b58acd7307156c)

### [Cyber Unicorns Reverse Engineered (CSPM & XDR): JSON, Cypher, Graphs, Embeddings, and Intruders](https://ai-blog.dev/cyber-unicorns-reverse-engineered-cspm-and-xdr-json-cypher-graphs-embeddings-and-intruders?utm_source=hashnode&utm_medium=feed)

This is a technical manifest of the data complexities involved in building AI XDR/CSPM Products in Cybersecurity

8 min read · [ai-blog.dev](https://ai-blog.dev/cyber-unicorns-reverse-engineered-cspm-and-xdr-json-cypher-graphs-embeddings-and-intruders?utm_source=hashnode&utm_medium=feed)

[Mar 18, 2025](https://hashnode.com/posts/cyber-unicorns-reverse-engineered-cspm-and-xdr-json-cypher-graphs-embeddings-and-intruders/67d9698dd354cd22f30da8dd)

### [RAG Agnostic Agentic AI: The OmniScience Design Pattern](https://ai-blog.dev/rag-agnostic-agentic-ai-the-omniscience-design-pattern?utm_source=hashnode&utm_medium=feed)

This article expands the limits of AI Agent Implementation by suggesting a new Design Pattern we all need to build a real RAG Agnostic AI Agent

8 min read · [ai-blog.dev](https://ai-blog.dev/rag-agnostic-agentic-ai-the-omniscience-design-pattern?utm_source=hashnode&utm_medium=feed)

[Feb 17, 2025](https://hashnode.com/posts/rag-agnostic-agentic-ai-the-omniscience-design-pattern/67b3a66a7c721f0bb370c062)
