AI Architect (AI for Security)
Neurons Lab · Remote · Romania · Slovenia · Germany · Spain Madrid · Hungary · Spain · Georgia · Slovakia · Latvia · Lithuania · Greece · Albania · Poland · Czech Republic · Spain Valencia · Moldova · Bulgaria · Estonia · Macedonia ·
COMPENSATION
€•• – ••k
Listed on the employer's job board (Ashby)
- Part-time (0.25-0.5 FTE)
- Fully remote
- Offensive Security + LLMs
- AWS Cloud Security
- Consulting Role
Join Neurons Lab as the offensive-security domain lead for a major iGaming client. You will partner with their advanced security team to challenge and improve their self-built AI exploitation pipeline. The work focuses on hardening LLM agents, optimizing model costs, and defining automated response strategies within strict regulatory constraints.
What you'll do
- Harden the end-to-end AI offensive pipeline from recon to sandboxed execution
- Design LLM agents for attack path planning and exploit validation
- Benchmark local open models against frontier models for cost and accuracy
- Define intrusion patterns and design automated response mechanisms
- Develop quick-win proofs of concept such as dependency scanning
- Present technical roadmaps and methodology to CISO and CTO stakeholders
- Ensure all sensitive data remains in-perimeter to meet compliance rules
What you'll need
- At least 3 years of hands-on offensive security or red-team experience
- Proven experience building LLM-driven security automation agents
- Expertise in vulnerability research and exploit development
- Experience running and tuning self-hosted open models on GPU
- Strong cloud security background, preferably with AWS
- Ability to lead technical discussions with executive security leaders
- Proficiency in Python and shell scripting
Nice to have
- Familiarity with iGaming regulations and certification constraints
- Experience with defensive security tools like SIEM and incident response
- Knowledge of underground exploit sources and threat intelligence
Stack
- Python
- AWS
- Nmap
- Metasploit
- Burp Suite
- Kali Linux
- Nuclei
- Kimi
- GPT-OSS
- MiniMax
- DeepSeek
How you'll work
- Initial allocation of 0.25 to 0.5 FTE
- Remote work arrangement
- Scaling potential as the security program expands