Available for Research & Consulting

AUSTRIA

AI SECURITY ENGINEER & RESEARCHER

Manuela Schrittwieser

Specializing in AI Engineering, AI Security Engineering, Technical Writing, and Security Research. I focus on securing Large Language Models, threat analysis for autonomous agents, and building resilient AI architectures.

Core Technical Expertise

LLM Security Threat Analysis Secure Development Penetration Testing Cloud Security AI Agent Security Security Research Technical Writing

Get in Touch Explore Projects

Featured Engineering & Research Projects

Open source tools, research frameworks, and security prototypes.

[ ]

SECURITY TOOL

LLM-Leak-Detector

Sentinel-LLM V3 is a multi-layered security guardrail that protects LLM applications on both sides of the conversation — scanning user prompts for injection attacks and model outputs for sensitive data leakage.

LLM-Security Guardrails Presidio

View Project GitHub

TECHNICAL ASSISTANT DEMO

Autonomous-SQL-Agent

A Technical Assistant Agent that translates natural language into SQL and executes it on a database. Trained with QLoRA on Qwen-2.5-1.5B.

Generative AI LLM Fine-Tuning SQL Generation

View Project 🤗 Hugging Face

INVOICE AGENT SERVICE

Structured_Data_Extractor

This repository contains an end‑to‑end pipeline for training, deploying and consuming a LLM‑based invoice extractor.

TextToJSON ServerlessInference Unsloth

View Project GitHub

HASHNODE PUBLICATION

NeuralStack | MS — Technical Blog

Exploring the intersection of artificial intelligence and cybersecurity. Deep dives into LLM vulnerabilities, AI security engineering, penetration testing neural models, and secure systemic development.

JULY 2026 Hashnode

The Centaur Defense Model – Cooperative Defense Strategies Between Humans and AI

This article analyzes the “Centaur Defense Model,” an advanced security concept in which humans and artificial intelligence (AI) form a united front against malicious or infected systems. The core of this model lies in moving beyond static firewall concepts in favor of a dynamic partnership that combines the computational speed and pattern recognition of AI with human ethical judgment and intuition.

Read Article

AUGUST 2026 Hashnode

Four AI Security Projects You Can Ship in a Month

AI security is a field where the credentialing infrastructure – certifications, courses, CTF badges – has outpaced the tooling infrastructure. Anyone can accumulate OWASP LLM Top 10 flashcard knowledge in a weekend. Far fewer people can point to a working prompt-injection detector, a scanner that finds real vulnerabilities in real MCP servers, or a red-team harness that produces an audit report a client could act on.

Read Article

SEPTEMBER 2026 Hashnode

Adversarial Attacks on Multi-Modal LLMs: How Image, Audio, and Video Inputs Get Weaponized and How to Defend Against It

This article maps the current attack surface across image, audio, and video inputs to vision-language and audio-language models, and lays out the defensive patterns that are actually holding up in practice, at the input layer, the model layer, and the system architecture layer.

Read Article

Want to read more technical articles?

Visit the full publication platform at neuralstackms.tech

Visit NeuralStack Blog

Let's Connect

Open for consulting, technical writing inquiries, research collaborations, and speaking opportunities.

Direct Email

ms.neuralstack@gmail.com

ALL LINKS

Linktree Profile

INSTAGRAM

@NEURALSTACK_MS

Social Channels:

LEGAL NOTICE / IMPRESSUM

Information according to Austrian Law

Media Owner / Creator: Manuela Schrittwieser

Profession / Field: AI Security Engineering

Location: Austria

E-Mail Contact: ms.neuralstack@gmail.com

Primary Domain: neuralstackms.tech

Disclaimer: Content published on NeuralStack | MS reflects independent technical research and analysis. All product and company names are trademarks of their respective holders.