🧠 DFG X-AI Cluster — Strategic Dashboard

Alessandro Golkar · Marcello Romano · Elena Steinvorth (AI moderator)
Last updated: 12 Mar 2026, 09:00 CET
Auto-synced from WhatsApp group

Knowledge Graph

Core Team
Potential Partners
Research Pillars
Institutions / Funding

Executive Summary

Alessandro and Marcello are exploring a DFG Cluster of Excellence proposal centered on autonomous intelligence for space systems. The core idea: future space systems cannot be teleoperated from Earth. Whether it's orbital robots, self-reconfiguring constellations, life support, or traffic management, the common thread is engineering systems that must function without human-in-the-loop, not by choice but by physical constraint.


The team wants to leverage agentic AI while staying grounded in their aerospace track record. Key tension: being ambitious enough for DFG while remaining credible as an aerospace-rooted team. A meeting is scheduled for Thu 19 March (12:30-15:30) at Alessandro's office to advance the proposal.

The Big Question (Evolving)

Current Working Formulation
"How do we engineer space systems that autonomously adapt, coordinate, and evolve in environments where real-time human control is physically impossible?"
Decision log:
Option A: Multi-agent systems in space (too standard, discussed for years)
Option B: Provably safe AI with space as testbed (too AI-centric, away from team's credibility)
Option C (current): Engineering autonomy under physical constraints. Bridges aerospace depth with AI ambition.
Alessandro's guidance: Favor scientific depth over application breadth. DFG is community-driven and academic. Must leverage existing track record while being ambitious with agentic AI. Think like German academics (conservative rigor).

Core Team

Potential Partners & Institutions

🔍 Gap identified: Team needs stronger AI/CS partners beyond Kochdumper. Jörg Ott fills networking, but core ML/agentic AI expertise still missing from the roster.

Open Points

Recommended Actions

🗓 Thu 19 March meeting: Use this to converge on the big question and identify 2-3 concrete research pillars. Come with a shortlist of external partners to approach.
👥 Alessandro & Marcello: Each write 3-5 bullet points on "what unique scientific contribution does our chair bring to this cluster?" before Thursday. This sharpens the positioning.
🤖 Elena: Prepare the Perplexity deep research prompt (DFG landscape, competing clusters, partner mapping). Ready to send on request.
🔬 Partner search: Approach Prof. Jörg Ott (TUM) for a preliminary conversation. Also explore TUM AI chairs (e.g., Günnemann, Cremers, Lienhart) and external universities (Stuttgart, Bremen, Darmstadt for aerospace; RWTH, KIT for AI).
📋 Marcello: Share initial thoughts on the cluster direction in the WhatsApp group. The more context before Thursday, the better the structured document Elena can prepare.

Suggested Explorations

💡 Dual-use framing: DFG values societal impact. The "autonomy under physical constraints" thesis applies beyond space (deep sea, disaster zones, remote infrastructure). This broadens the appeal without diluting the core.
💡 Industry advisory board: The Villa Vigoni network (Isar Aerospace, D-Orbit, OHB, OroraTech, Alpine Space Ventures) could form a natural industry advisory board for the cluster. Synergy between both initiatives.
💡 Kochdumper as bridge: His formal verification background (Althoff group / CORA) connects to the broader safe AI community. Could open doors to AI safety researchers who wouldn't normally engage with aerospace.
💡 Timing strategy: The DFG Excellence Strategy round 3 hasn't been announced. Use this pre-announcement period to build alliances and publish joint papers. By the time the call opens, the consortium should already be working together.