AI Architect (AI for Security)

neurons-lab.comRomaniaAI Engineering

Apply on company site

Opens the official neurons-lab.com careers page

Salary

Not listed

Job type

Full-time

Work mode

Remote

Experience

Senior · 3+ yrs

Posted

11 Sept 2026

Status

Open · checked 7d ago

About this role

This role suits an experienced offensive security specialist based in Romania who wants to lead AI-driven security architecture engagements for a regulated gaming client. The position involves evaluating and improving automated exploitation and anomaly detection systems through hands-on technical consulting.

Role Overview

Neurons Lab is seeking an AI Architect specializing in AI for Security for a position based in Romania. The professional will operate as the security domain owner, engaging directly with a client's chief information security officer team to challenge, refine, and optimize an advanced artificial intelligence-driven offensive security capability and runtime anomaly detection layer.

Key Responsibilities

Participate in joint working sessions with client security engineers to assess and harden the end-to-end artificial intelligence-driven offensive pipeline.

Design and refine exploitation agents, focusing on how language models plan attack paths, select and validate exploits, and orchestrate parallel sandboxes.

Benchmark local and open-source models against frontier models to optimize the cost-per-finding of existing pipelines.

Shape the runtime anomaly detection layer by defining intrusion and privilege-escalation precursor patterns, automated responses, and triage routing.

Develop quick-win proof-of-concept projects, such as vulnerability scanning passes or head-to-head model benchmarks.

Translate technical findings into structured proposals and roadmaps for technical leadership.

Maintain strict data sovereignty and in-perimeter build-time constraints without utilizing external model providers.

Requirements

Three or more years of hands-on experience in offensive security, vulnerability research, or red-teaming.

Demonstrable background in exploit development, chaining, and threat intelligence.

Experience integrating large language models into security workflows such as reconnaissance and exploitation.

Practical knowledge of running self-hosted and local open models with a focus on hardware and cost.

Capability to lead technical conversations with security executives and act as the primary domain expert.

Cloud hyperscaler experience, with a preference for Amazon Web Services.

Proficiency in Python and shell scripting.

Nice to Have

Background in the iGaming or regulated infrastructure sector and associated certification constraints.

Defensive security experience including security information and event management, anomaly detection, and incident response.

Skills and Tools

Penetration testing tools including Nmap, Nuclei, Katana, Acunetix, Metasploit, Burp Suite, and Kali Linux.

Large language model agent orchestration, prompt design, and guardrails.

Self-hosted open models such as Kimi, GPT-OSS, MiniMax, and DeepSeek.

Cloud security, container isolation, and secure inference hosting on Amazon Web Services.

Python and shell programming.

About the Company

Neurons Lab provides artificial intelligence architecture and consulting services. The organization partners with external clients on specialized technical engagements, delivering domain expertise and methodology oversight through structured consulting teams.

How to Apply

Interested candidates should visit the official Neurons Lab careers page to submit an application for this opportunity.

Skills and tech stack

Python Shell Nmap Nuclei Katana Acunetix Metasploit Burp Suite Kali AWS Kimi GPT-OSS MiniMax DeepSeek AWS Bedrock LLM LLMs

Questions candidates ask

What is the primary objective of this role?

The primary objective is to act as the security domain owner, challenging, refining, and optimizing an advanced artificial intelligence-driven offensive security pipeline and runtime anomaly detection layer for a regulated client.

Where is the position located?

The position is based in Romania.

What level of experience is required in offensive security?

Candidates must have three or more years of hands-on experience in offensive security, vulnerability research, or red-team operations.

What programming languages are needed?

The professional needs to write their own code using Python and shell scripting.

What cloud platform is preferred?

Amazon Web Services is the preferred cloud hyperscaler experience for this position.

Ready to apply?

The application is on the official neurons-lab.com careers page.

Apply on company site