Kutató – DeepRAP Challenge

What the Challenge is funding: Research that moves beyond today's deep learning/RL paradigms toward genuinely trustworthy cognitive AI; causal reasoning, abstraction, and planning under uncertainty. Funded projects don't just publish a paper, they build and demonstrate a working system (TRL3/4), help define new industry benchmarks for reasoning and trustworthiness, and join a wider EU-funded portfolio shaping how cognitive AI gets built and regulated across Europe. The opportunity: Visibility at the frontier : this is exactly the kind of work that gets cited, gets you invited to speak, and gets you noticed by labs and industry doing serious reasoning/planning research A funded system, not a thought experiment : you're not writing a proposal that sits in a drawer; a successful award means building and demonstrating the actual architecture Named contributor on a flagship EU AI initiative: tied to benchmark development and portfolio activities the EIC is running across all funded DeepRAP projects A credential that compounds: EIC Pathfinder co-authorship is a strong signal on any postdoc, faculty, or industry research application going forward What you'd do: Co-develop the scientific narrative and technical approach for a 30-page Pathfinder proposal Bring rigor on reasoning/abstraction/planning methodology, related work, and evaluation design Work directly with our founding team through submission Who we're looking for: Research background in neuro-symbolic AI, causal inference, cognitive architectures, or deep RL/planning Track record: publications at NeurIPS/ICML/ICLR/AAAI or equivalent, PhD in progress or completed Based at or affiliated with a top research institution (ETH Zurich, Imperial College, EPFL, Oxford, TU Munich, etc.) EU/associated-country affiliation strongly preferred Comfortable working fast, iteratively, with a startup team, not academic-committee pace Originally posted on Himalayas