AI Safety Researcher
We’re looking for an experienced AI Safety Researcher to support safety, trust, and alignment work across large‑scale AI systems. You’ll collaborate with research, engineering, and T&S teams to run adversarial tests, improve system alignment, and produce high‑quality training and evaluation data.
Responsibilities Include:
1. Lead adversarial testing/red‑teaming to identify safety gaps.
2. Work with LLM evaluation tools, automated red‑teaming frameworks, and large datasets.
3. Improve system alignment through prompt engineering, preference tuning, and prompt optimization.
4. Benchmark safety performance and support scalable safety processes.
Skills Required:
5. Demonstrated experience in AI safety, adversarial testing, or system alignment.
6. Strong programming skills: Python, Java, SQL.
7. Hands‑on experience with LLMs and prompt/context engineering.
8. MSc/PhD (completed or in progress) in AI/ML or related field preferred.
9. Bonus: experience with cross‑language models.
Please note this is a contract position, Inside IR35 and fully remote
Job Title: AI Safety Researcher
Location: London, UK
Job Type: Contract
Trading as TEKsystems. Allegis Group Limited, Maxis 2, Western Road, Bracknell, RG12 1RT, United Kingdom. No. 2876353. Allegis Group Limited operates as an Employment Business and Employment Agency as set out in the Conduct of Employment Agencies and Employment Businesses Regulations 2003. TEKsystems is a company within the Allegis Group network of companies (collectively referred to as "Allegis Group"). Aerotek, Aston Carter, EASi, Talentis Solutions, TEKsystems, Stamford Consultants and The Stamford Group are Allegis Group brands.