PhD position attacks against language models; LLMs
Job in
7500, Enschede, Overijssel, Netherlands
Listing for:
Rotterdam School of Management, Erasmus University (RSM)
Full Time
position
Listed on 2026-01-23
Job specializations:
-
IT/Tech
Data Scientist, AI Engineer, Data Analyst, Machine Learning/ ML Engineer
-
Research/Development
Data Scientist
Salary/Wage Range or Industry Benchmark: 60000 - 80000 EUR Yearly
EUR
60000.00
80000.00
YEAR
Job Description & How to Apply Below
Position: PhD position on attacks against large language models (LLMs)
Overview
This project will investigate attacks on large language models (LLMs), a major recent development in artificial intelligence that has already seen many integrations into public life. If these LLMs can be triggered into providing malicious output, this may have disastrous consequences, leading to the generation of harmful content, the execution of malicious code on connected devices, or the abuse of limited resources.
The idea is to assess the resistance of these models against new attacks, using techniques coming from the domain of AI and optimisation, and develop methods to defend against such harms by leveraging cryptographic approaches.
Responsibilities
Investigate how to adapt existing adversarial attacks for image classification and other domains against existing open source LLM (e.g., Llama 3, Phi-3), as well as develop new kinds of attacks, for example based on evolutionary algorithms.Investigate to what extent data poisoning attacks can influence the output of LLM models in security and safety critical infrastructure.Perform the attack under different scenarios and model the impact.Evaluate the impact of such attacks when executed in multi-agentic systems, where the output of one LLM is used as input for another LLM.Design new defense methods, e.g., inspired by cryptography, to prevent such attacks from affecting real-world LLM systems, where we focus on methods that limit the computational overhead to minimize the energy, and therefore environmental cost of such defences.These research directions will advance the understanding of security vulnerabilities of LLMs, and the prevention of malicious output generation.
#J-18808-Ljbffr
Note that applications are not being accepted from your jurisdiction for this job currently via this jobsite. Candidate preferences are the decision of the Employer or Recruiting Agent, and are controlled by them alone.
To Search, View & Apply for jobs on this site that accept applications from your location or country, tap here to make a Search:
Search for further Jobs Here:
