PhD position: Multimodal Generative AI for Embodied Collaborative Agents

PhD Position: Multimodal Generative AI for Embodied Collaborative Agents

Location: Edinburgh, United Kingdom, Interaction Lab and National Robotarium , Heriot-Watt University

Collaboration with Toshiba Europe Limited, Cambridge Research Laboratory, UK.

Duration: 3.5 years, start date : between October 2025 - Jan 2026

Note : eligible for "home" students only, see definition [here]

Project Overview:

We are seeking a highly motivated PhD candidate to contribute to an innovative research project focused on developing multimodal generative AI for embodied collaborative agents. This project aims to create AI agents capable of seamlessly collaborating with humans and other agents in dynamic environments, utilizing Vision, Language, and Action (VLA) models. The research will explore ad-hoc teamwork, continual learning, and multimodal conversational  communication, with applications in healthcare and retail domains. The research will be conducted at the UK National Robotarium labs and in collaboration with Toshiba (Cambridge, UK).

Research Focus:

• Developing robust frameworks for multi-agent ad-hoc teamwork.

• Integrating generative AI Vision, Language, and Action (VLA) models for natural, multimodal, conversational  collaboration.

• Enabling agents to understand visual cues, interpret and generate collaborative natural language, and execute coordinated actions.

• Investigating continual learning and adaptation in novel environments.

• Exploring efficient and transparent task decomposition, negotiation, and role allocation within agent teams.

Methodology:

The project will leverage state-of-the-art generative AI techniques, continual learning, and meta-learning approaches. Simulated environments and real-world data collection at the UK National Robotarium will be used for training and evaluation.

Application Domains:

The research will be evaluated in healthcare and retail settings, focusing on improving efficiency, reducing errors, and enhancing agent collaboration.

Requirements:

• A strong background in computer science, artificial intelligence, NLP, machine learning, or a related field.

• Experience with Python and other relevant programming languages.

• Interest in multimodal interaction and robotics.

• Excellent communication and teamwork skills.

Contact:

For more information, please contact Dr Alessandro Suglia (a.suglia@hw.ac.uk) and Professor Oliver Lemon (o.lemon@hw.ac.uk) 

How to Apply:

Please include a CV, cover letter, and any relevant supporting documents.