About Us:
At Persistent Security Industries, we build AI models that plan and execute attacks against real security controls and consult defenders about which defenses work and how to apply them! Our core platform, Nemesis, runs offensive techniques against real live infrastructure, evaluates the results, and adapts in real time.
We have a working product with a growing number of customers, join us to take a system that already ships and make it measurably better, then get it into customer environments: local, secure and observable.

The Shape of the Role:
The role sits at the intersection of offensive security and applied ML. You need to have done the security work yourself: discovered vulnerabilities, written exploits, evaded detections and gave crucial advice to defenders.The ML engineering you can pick up on a mature codebase with a team that knows the domain. You're bringing the judgment to guide model-generated attacks and tell whether they are realistic or nonsense.

What You‘re Working With:
You work on he full loop from data to a served model to using it for R&D: sourcing and curating training data, quality and privacy gates, fine-tuning, alignment, deployment. It works and ships models today. Your job is to make each stage better, faster, and more measurable, and to own what goes into the corpus and why.
We can push past standard synthetic training methods: With Nemesis the signal is whether a real attack actually succeeded against real defenses. This is one of the biggest levers for our success!
All of that enables us to deploy models in customer environments: on-premise, air-gapped, compliant on custom hardware, adapted to their stack.

Experience You Should Have:

• Significant hands-on offensive security experience: penetration testing, red teaming, adversary simulation, or security research. You have operated against real defenses, not just read about them.

• Strong software engineering in Python. Comfortable with distributed systems, API design, databases, and cloud infrastructure.

• Experience building production systems on top of LLMs: API integration, prompt engineering, structured output, tool use.

• Based in the EU, or willing to work EU-compatible hours.

• Fluent English. We are an international team.

What We Offer

• Remote-first, diverse team of great people

• Direct impact, your lab is the real tech stack, in the wild, so you know if you make impact!

• A mature product with paying customers to work on

• Dedicated infrastructure for training and serving

How to Apply
Send a short note on the most interesting attack you have run or defense you have broken, and how you would want to teach a model about it along with your resume to info@persistent-security.net