Sameer Ghanima

Sameer Ghanima

Founder at Sentinely, AI agent security

About

Building Sentinely, a runtime security layer for AI agents. We stop prompt injection, memory poisoning, and agent drift before damage is done. Previously built AI products for clients watched too many agents get manipulated in production, so I built the tool I wished existed.

Badges

Tastemaker
Tastemaker
Gone streaking
Gone streaking

Maker History

  • The Smartest
    The SmartestChallenge Your Knowledge and Win Big on TheSmartest
    Jun 2024
  • 🎉
    Joined Product HuntMay 24th, 2024

Forums

Sentinely - Runtime security for AI agents. 3 lines integration.

Most AI agent security tools filter inputs before they reach the model. Sentinely works inside the pipeline, scoring every action before it executes. It detects prompt injection, tracks each agent's behavioral baseline and flags drift, quarantines suspicious memory writes, and catches multi-agent manipulation. Every action is logged with a full audit trail for SOC2 and EU AI Act compliance. 3-line integration. Works with LangChain, CrewAI, OpenAI, and Vercel AI SDK.
View more