LLM Security Intern


Job Location:

Tübingen - Germany

Monthly Salary: Not Disclosed
Posted on: Yesterday
Vacancies: 1 Vacancy

Job Summary

AI agents are being deployed everywhere making high-stakes decisions and increasingly automating research itself. Yet their reliability and security remain unsolved technical problems on a frontier that keeps shifting. Our mission at Exponential Security Labs is to secure agentic systems through self-improving red-teaming and guardrail agents leveraging autonomous AI research. We sit at the intersection of safety and self-improvement building on a decade of research in adversarial robustness and AI safety. Were looking for exceptional people to join us!

Tasks

As an LLM Security Intern you will help us design and build a scalable ML system from the ground up. This includes:

  • Build and maintain production ML/LLM systems
  • Deploy optimize and monitor inference pipelines
  • Evaluate models build red-teaming agents and guardrails
  • Improve reliability latency and cost of model serving
  • Shape best practices around AI safety and reliability

Requirements

What we are looking for:

  • MSc or PhD in ML NLP computer science or related field
  • Experience with LLMs and agentic systems
  • Strong Python and ML engineering fundamentals
  • High ownership autonomy bias toward execution

Benefits

We offer:

  • Path to a full-time position
  • Competitive compensation
  • Flexible hours remote-friendly setup
  • Top-tier equipment and tooling
  • Growing team low bureaucracy direct impact
  • A chance to safeguard the agentic playground
AI agents are being deployed everywhere making high-stakes decisions and increasingly automating research itself. Yet their reliability and security remain unsolved technical problems on a frontier that keeps shifting. Our mission at Exponential Security Labs is to secure agentic systems through sel...

About Company

Company Logo

We are a recently founded startup on agentic AI security based in Tübingen Germany.

View Profile View Profile