Alignment Red Team - Research Engineer/Research Scientist
We are seeking an Alignment Red Team Research Engineer/Research Scientist to investigate how advanced AI systems can fail, be misused, or behave unexpectedly under challenging conditions. You will design and conduct rigorous adversarial evaluations of models, agents, and safety mechanisms, with the goal of identifying vulnerabilities before they create meaningful real-world risks. The role combines scientific research, engineering, critical thinking, and practical experimentation in a highly collaborative environment.
Your responsibilities will include developing threat models and evaluation methodologies; creating adversarial prompts, scenarios, tools, and test environments; probing models for deceptive, unsafe, biased, or otherwise misaligned behaviour; and building automated pipelines to support large-scale red-team exercises. You will analyse results, distinguish genuine failures from artefacts, document reproducible findings, and communicate their significance to technical and non-technical audiences. You may also contribute to mitigations, safety benchmarks, model monitoring approaches, and research publications or internal reports. The work may involve exploring capability elicitation, agentic behaviour, instruction following, reward hacking, situational awareness, scalable oversight, and security or privacy risks.
We are looking for candidates with experience in machine learning, software engineering, AI safety, cybersecurity, formal methods, or a closely related field. Strong programming skills, particularly in Python, and the ability to work effectively with modern AI models and evaluation tooling are expected. You should be comfortable designing experiments under uncertainty, reasoning about adversarial behaviour, and turning ambiguous questions into clear, testable investigations. Research experience demonstrated through publications, substantial independent projects, or equivalent practical work is valuable, as is familiarity with language models, reinforcement learning, agent systems, or security testing. Curiosity, intellectual honesty, careful documentation, and the ability to collaborate across disciplines are essential. Candidates from non-traditional backgrounds are encouraged to apply if they can demonstrate relevant expertise and a strong commitment to improving the safety and reliability of advanced AI systems.
Alignment Red Team - Research Engineer/Research Scientist
Other similar jobs
Popular job searches
Your next job
starts here.
JOB SPECIALISMS
LATEST JOBS
TOP SEARCHES
LOCATIONS
- Security Engineer
- Security Analyst
- Security Architect
- IT Security Manager
- Security Consultant
- SOC Analyst
- Identity Access Management IAM
- Cyber Security Consultant
- Cloud Security
- Application Security
- Incident Response
- Penetration Tester
LATEST JOBS
- Security Engineer (Contractor)
- Vice President, Field CISO Com...
- Microsoft 365 & Security Infra...
- Senior ICT Security Designer
- Data Protection Expert
- Data Protection Officer: Schoo...
- Data Protection Officer
- IT CONSULTANT (IT Management,...
- Technology and Cyber Employee...
- Cyber Security Consultant
- Cyber Security Consultant (Pen...
- Platform and Services Engineer...