Postdoc in LLM-based Decision Support for Multimodal Situation Awareness

About the role

Two-year postdoctoral position at Chalmers in Gothenburg, building LLM-based decision support systems that fuse multimodal sensor streams (vision, lidar, audio, text) for real-time situation awareness. The host group collaborates with Swedish defence and emergency-response partners. Deadline 1 June 2026.

Responsibilities

• Build and fine-tune LLM-based pipelines that consume multimodal inputs and produce structured recommendations.
• Design retrieval, planning, and grounding layers that keep model outputs verifiable against sensor evidence.
• Publish in NeurIPS, ICML, EMNLP, and IROS or similar.
• Run user studies with operational partners to evaluate decision quality and trust.

Requirements

• Doctoral degree in computer science, AI, or a related field by the start date.
• Hands-on experience with modern LLMs (fine-tuning, RAG, agents) and at least one multimodal stack.
• Python and PyTorch proficiency.
• Fluent English.

Nice to have

• Prior work on tool use, planning, or verifiable LLM outputs.
• Background in autonomous systems, robotics, or human-AI teaming.
• Publications at top ML venues.

How to apply

Open the apply link and submit CV, publication list, research statement, and two references via the Chalmers recruitment system before 1 June 2026.