This thesis proposes a factual sustainability moderator based on small AI/ML reasoning models, NLP, knowledge graph reasoning concepts, or similar, depending on the choice of the students. It would include testing the proposed moderation features to evaluate their helpfulness, accuracy, and overall workshop quality experienced with and without them, in an iterative manner, based on previous research into moderation techniques.
The research question in focus is how an AI agent can interrupt and be helpful in a workshop setting, producing relevant questions to the workshop participants in a non-intrusive, polite way – as a respected moderator?
About the company
Sustainable AI Solutions, SAIS, is a new, ground-breaking Swedish AI and sustainability start-up developing sustainability solutions based on AI technologies. Their goal is to accelerate the transformation of sustainability with digitalization and AI. SAIS is still in the beginning stages. Join them now and contribute to the exciting journey.
The role
We are looking for two thesis workers with a passion for AI/ML and sustainability. This is a unique opportunity to experience a start-up phase together with professional sustainability experts. This role will collaborate with other thesis workers and company employees.
Your profile
You have a solid background in data, computer science, system design, or an IT, programming, AI/ML education path. You have experience in using modern AI tools and know of both risks and opportunities.
You are a good team player and able to interact with friendly users and various researchers. You are eager to learn more, innovative, and scientific, with an interest in sustainability. You have an ethical and respectful approach to data and science.
Thesis details
There is a need to scale up fact-based sustainability actions from organizations, companies, and governments. This thesis is to select and trial different AI/ML components and technologies to help moderate sessions of an existing factual sustainability process and create a suite of moderation features to help end-users learn, evaluate, and optimize their business to net zero.
The technologies involved will partly be up to the student to propose, evaluate and compare from a human interaction, sustainability and moderation efficiency view. Candidate technologies that have been identified as possible building blocks are
- STT/TTS for real-time listening and intervention
- Reasoning using AI/ML over sustainability facts and discussion topics and summaries
- Adopting meeting moderation discussion techniques to create opportunities for consensus on facts, while creating opportunities for individual innovation proposals on solutions
- Based on existing sustainability research and a Factual example database being created
- STT/TTS, real-time conversational AI, faster-whisper, AI/ML reasoning models
Application
The application deadline is November 16th, 2025. Please submit your application, CV, grades, and a brief letter outlining your interest in the task and proposal.
We look forward to hearing from you!