The Agentic Systems Lab, part of the Chair of Information Management at ETH Zürich, is dedicated to designing, developing, and evaluating agentic AI systems in real-world environments.

We are committed to helping ambitious, high-agency students grow into the next generation of AI talent by:

🔗 Connecting them with a community of like-minded builders

🚀 Providing access to leading AI startups and innovators

🏭 Curating opportunities to apply agentic AI in established industries

🎓 Offering hands-on guidance and cross-sector expertise from researchers and industry collaborators

Interested in a thesis or semester project? Join our Genesis Batch 2025, a collaborative cohort of 10–20 master’s students working on agentic AI applications with real-world impact.

• Lab Track – Collaborate on internal lab projects focused on agentic system design, development, and evaluation.
• Explorer Track – Partner with startups or industry collaborators to ideate, prototype, and validate agentic systems.
• Startup Track (Bring Your Own Project) – Develop your venture or prototype into a robust agentic application or publication-ready research paper.

We’re actively designing and evaluating agentic AI across several domains:

• Agentic AI in Education: Developing interactive, multi-level assessment platforms using AI-powered chat and voice examinations.

• Agentic AI in Healthcare: Exploring personalized AI-driven nutrition guidance and meal planning, tailored to health data and wellness goals.

• Agentic AI in Science: Building an AI-augmented manuscript review system, providing multi-AI feedback to strengthen academic submissions.

• Agentic AI in Insurance: Investigating AI-assisted claims processing, personalized policy purchasing, and fraud detection.

• Time Series Language Models (TSLMs): Launching open-source TSLMs for reasoning over and predicting temporal data.

• Behavioral Economics of LLMs: Studying decision-making biases in human–AI and AI–AI interactions through behavioral experiments.

• Browser Use (YC W25) is building an open-source platform that allows AI agents and users to automate web tasks without writing complex code.
• nunu.ai (a16z and YC W23) is building the first multimodal AI agents that can autonomously play and test video games, revolutionizing quality assurance in game development without manual scripting.
• Bloom (YC X25) is combining conversational AI with instant app deployment, empowering creators, developers, and entrepreneurs to turn ideas into real apps in seconds.
• Taskbase is developing AI-powered tools to make learning more personalized, effective, and scalable.

Students looking to get involved can apply to one of our tracks using form.

Our lab welcomes collaboration with industry, startups and research groups. Interested parties can reach us at agenticsystems@ethz.ch or through our channels below.

• Email: agenticsystems@ethz.ch
• Website: agenticsystems.ethz.ch
• Discord: Join our community
• Twitter/X: @ETH_agent_lab
• LinkedIn: Agentic Systems Lab