Stefan Schmid
Professor
TU Berlin
From Topology-Aware to Fabric-Adaptive: Rethinking Datacenter Networks for AI
AI has become a systems problem. Training and serving today's models is limited less by algorithms than by the machines and networks underneath them: interconnect topology, congestion, scheduling, memory and energy.
NetSysAI brings together researchers and industry engineers working at that boundary, from datacenter fabrics and self-adjusting networks to efficient models and hardware-aware learning, for a day of talks and discussion in the Negev.
Stefan Schmid
Professor
TU Berlin
From Topology-Aware to Fabric-Adaptive: Rethinking Datacenter Networks for AI
Eitan Zahavi
Distinguished Architect
NVIDIA
Talk title TBA
Asaf Aharoni
Distinguished Software Engineer
Talk title TBA
Noam Attias
Director of Engineering
Amazon · Annapurna Labs
Talk title TBA
Chaim Baskin
Assistant Professor | Fundamental Research Lead
BGU · Decart
What to Keep and How to Remember: Towards Efficient World Models
Mark Silberstein
Professor
Technion
Talk title TBA
Stefan Schmid is a Professor at the Technical University of Berlin, Germany. MSc and PhD at ETH Zurich, Postdoc at TU Munich and University of Paderborn, Senior Research Scientist at T-Labs in Berlin, Associate Professor at Aalborg University, Denmark, Full Professor at the University of Vienna, Austria, and sabbatical as a Fellow at the Israel Institute for Advanced Studies (IIAS), Israel. Stefan Schmid received the IEEE Communications Society ITC Early Career Award 2016, an ERC Consolidator Grant 2019, and an ERC Proof-of-Concept Grant 2026. He is the Editor in Chief of the EATCS Bulletin, was a co-chair of ACM Sigmetrics and ACM MobiHoc, and served as local organizer of DISC and IEEE NetSoft.
Tentative The detailed program with talk titles will be posted here. Times are indicative and subject to change.
| 09:30 | Registration, coffee & networking |
| 10:00 | Welcome, greetings and opening remarks BGU, DAAD and partners · Chen Avin (BGU) |
| 10:30 | Morning session Keynote, Stefan Schmid (TU Berlin): “From Topology-Aware to Fabric-Adaptive: Rethinking Datacenter Networks for AI” Two invited talks, speakers and titles TBA |
| 12:45 | Lunch break DAAD information stand for short consultations |
| 13:30 | Afternoon session Three invited talks, speakers and titles TBA |
| 15:00 | Coffee break DAAD information stand for short consultations |
| 15:30 | Graduate student session Short talks by MSc and PhD students |
| 16:30 | Closing remarks |
A dedicated block of short talks by MSc and PhD students. If you are working on networking or systems for AI, we invite you to present your work in a short talk (about 10–15 minutes).
Email the organizers at avin@bgu.ac.il by 31 October 2026 with your talk title, a short abstract and links to any relevant publications. Applying for a talk is separate from registration, so please register as well.
Attendance is free, but registration is required (for campus entry, catering and seating). The form takes about a minute; if you are coming by car, it also asks for your car details for the campus gate.
Marcus Family Campus, Be'er Sheva, Israel.
Building and hall: TBA. Signs will be posted from the campus entrances.
Campus maps: bgu.ac.il · Train schedules: Israel Railways · Buses: Moovit or Google Maps.
Take the train to Be'er Sheva North / University station (not Be'er Sheva Center). The station connects directly to campus via a covered pedestrian bridge, a few minutes' walk. Direct trains run from Tel Aviv, Haifa and other cities along the coast.
Take the train from the airport station toward Tel Aviv, change at Tel Aviv HaHagana to a Be'er Sheva train, and get off at Be'er Sheva North / University. By car or taxi the trip takes about 90 minutes, depending on traffic.
From the center of Israel, take Route 6 or Route 40 south to Be'er Sheva. Parking instructions and gate access for registered participants will be posted here closer to the date.
Please bring an ID for campus entry. The workshop is held on a Thursday; public transport runs normally.
Questions? Contact us at avin@bgu.ac.il.
