Research Associate on Rethinking Multi-agent Systems in the Era of LLMs
| Dyddiad hysbysebu: | 10 Mehefin 2025 |
|---|---|
| Cyflog: | £38,674 i £40,855 bob blwyddyn, pro rata |
| Oriau: | Llawn Amser |
| Dyddiad cau: | 04 Gorffennaf 2025 |
| Lleoliad: | Oxford, Oxfordshire |
| Gweithio o bell: | Ar y safle yn unig |
| Cwmni: | University of Oxford, Dept of Computer Science |
| Math o swydd: | Cytundeb |
| Cyfeirnod swydd: | 180134 |
Crynodeb
The Department of Computer Science seeks to employ 2 postdoctoral researchers to work on a new project in the area of LLMs/multi-agent systems, under the direction of PI Professor Michael Wooldridge. The posts are full-time, fixed term positions: Post One is available for 18 months; Post Two is available for 12 months. Both posts are tenable immediately, to conclude before May 31st 2027.
The multi-agent systems paradigm within the area of AI has been widely researched since the early 1990s, but has come to prominence recently with the emergence of Large Language Models (LLMs). The possibility of using networks of LLM-based agents to solve complex problems has shown great potential, for example in DeepMind’s Co-Scientist System. Against this background, the goal of this project is to reframe the field of multi-agent systems in the context of contemporary LLMs. If we are successful, then we will have constructed a new “Standard Model” for multi-agent systems: one which embraces LLMs, and leverages them to their maximum advantage. More specifically, we will develop both a methodology for deploying LLMs in individual agents (an “LLM-based agent architecture”) and a methodology for building LLM-based multi-agent systems for a range of typical interaction scenarios (negotiation, evaluation, etc). Our work is expected to embrace emerging standards for LLM agents (notably MCP, A2A).
Candidates will be expected to have a PhD (or be close to completion) in a related area and should indicate their special area(s) of interest from the four topics above in a cover letter. The primary selection criteria will be relevant research experience and publication track record.
What We Offer
As an employer, we genuinely care about our employees’ wellbeing and this is reflected in the range of benefits that we offer including:
• An excellent contributory pension scheme
• 38 days annual leave (pro-rata for part-time roles)
• A comprehensive range of childcare services
• Family leave schemes
• Cycle loan scheme
• Discounted bus travel and Season Ticket travel loans
• Membership to a variety of social and sports clubs
Diversity
Committed to equality and valuing diversity.
Application Process
You will be required to upload a supporting statement and an up-to-date CV as part of your online application.
Your supporting statement must clearly demonstrate how you meet each of the essential selection criteria listed in the job description. Applications that do not include a supporting statement or CV, or fail to address the criteria in sufficient detail, will not be considered.
While we recognise the value of AI tools in assisting with application preparation, submissions that are clearly AI-generated without personalisation or insight will be rejected. It's crucial that your application reflects your own experiences and understanding of the role.
The closing date for applications is noon on 4th July 2025. Interviews are expected to be held in July.
The multi-agent systems paradigm within the area of AI has been widely researched since the early 1990s, but has come to prominence recently with the emergence of Large Language Models (LLMs). The possibility of using networks of LLM-based agents to solve complex problems has shown great potential, for example in DeepMind’s Co-Scientist System. Against this background, the goal of this project is to reframe the field of multi-agent systems in the context of contemporary LLMs. If we are successful, then we will have constructed a new “Standard Model” for multi-agent systems: one which embraces LLMs, and leverages them to their maximum advantage. More specifically, we will develop both a methodology for deploying LLMs in individual agents (an “LLM-based agent architecture”) and a methodology for building LLM-based multi-agent systems for a range of typical interaction scenarios (negotiation, evaluation, etc). Our work is expected to embrace emerging standards for LLM agents (notably MCP, A2A).
Candidates will be expected to have a PhD (or be close to completion) in a related area and should indicate their special area(s) of interest from the four topics above in a cover letter. The primary selection criteria will be relevant research experience and publication track record.
What We Offer
As an employer, we genuinely care about our employees’ wellbeing and this is reflected in the range of benefits that we offer including:
• An excellent contributory pension scheme
• 38 days annual leave (pro-rata for part-time roles)
• A comprehensive range of childcare services
• Family leave schemes
• Cycle loan scheme
• Discounted bus travel and Season Ticket travel loans
• Membership to a variety of social and sports clubs
Diversity
Committed to equality and valuing diversity.
Application Process
You will be required to upload a supporting statement and an up-to-date CV as part of your online application.
Your supporting statement must clearly demonstrate how you meet each of the essential selection criteria listed in the job description. Applications that do not include a supporting statement or CV, or fail to address the criteria in sufficient detail, will not be considered.
While we recognise the value of AI tools in assisting with application preparation, submissions that are clearly AI-generated without personalisation or insight will be rejected. It's crucial that your application reflects your own experiences and understanding of the role.
The closing date for applications is noon on 4th July 2025. Interviews are expected to be held in July.