Affiliated with
Advancing 6G Photonic
Networks &
THz Wireless
01 — Latest News
Recent updates on publications, awards, and teaching.
Awarded the Best Student Paper Award at OECC 2026 for a record 221 Gb/s dual-polarization THz wireless transmission over 500 m in urban Berlin — a rate–distance product exceeding 100 Tb/s·m in the 300 GHz band.
View on LinkedIn →Paper accepted to IEEE Photonics Technology Letters (PTL): “Neural-Network Based Transceiver Nonlinearity Mitigation for Probabilistically Shaped Photonic THz Wireless Transmission.”
Started co-founding LUXONNIC as CEO — an early-stage venture-in-formation at Fraunhofer HHI developing photonic neural transceivers for AI data centers — and submitted a funding proposal to the BMFTR xG-Incubator programme (StartUpConnect).
Started as Lecturer at HTW Berlin, teaching the B.Sc. programme in Information and Communication Technology.
Two papers published in the Journal of Lightwave Technology (JLT).
02 — Research
From experimental 300 GHz THz links to neuromorphic photonics — building the physical layer of 6G.
Experimental 300 GHz (WR-3) transmission systems built on photonic upconversion — from PIN-photodiode transmitters to receiver-side DSP. Demonstrated throughputs beyond 109 Gbit/s, and a record 221 Gb/s dual-polarisation link over 500 m.
Rate-adaptive probabilistic shaping approaching the Shannon limit, integrated into converged fiber–THz backhaul architectures for 6G — the subject of my PhD dissertation at TU Berlin.
Optical signal processing on integrated photonic chips using neural networks — modulation format identification, signal correction, and feature extraction entirely in the optical domain at sub-nanosecond latency.
03 — Venture
Co-Founder & CEO of an early-stage deep-tech venture, developed at Fraunhofer HHI, working to reinvent the optical transceiver for AI data centers.
“Becoming the secret sauce of next-gen transceivers.”
Optical transceivers already consume roughly 10% of an AI factory’s power — about 40 MW per site, much of it spent on digital signal correction. LUXONNIC is developing the LUXONNICore, a photonic integrated circuit that performs signal equalization directly in the optical domain using photonic neural networks — replacing the power-hungry DSP chips inside next-generation (LPO/CPO) transceivers and targeting up to 30× higher bit-per-Joule efficiency.
04 — Publications
First-author papers highlighted. Google Scholar • IEEE Xplore
In-Ho Baek, J. Honecker, M. Bittiehn, R. Elschner, C. Schubert, R. Freund
In-Ho Baek, O. Stiewe, R. Elschner, M. Rösch, A. Tessmann, M. Nölle, L. Molle, C. Schubert, R. Freund
Best Student Paper Award -- OECC 2026
In-Ho Baek, O. Stiewe, J. Gläsel, S. Nellen, G. Schwanke, T. T. Tran, R. Elschner, C. Schubert, R. B. Kohlhaas, P. Runge, M. Schell, R. Freund
In-Ho Baek, O. Stiewe, J. Gläsel, M. F. Ulukan, A. Schindler, F. Ganzer, T. T. Tran
In-Ho Baek, O. Stiewe, S. Nellen, G. Schwanke, R. Elschner, C. Schubert
In-Ho Baek, F. Bart, R. Elschner, F. Meier, D. Hellmann, A. Maassen, C. Schubert
R. Elschner, F. Bart, F. Meier, O. Stiewe, IH Baek et al.
C. Schubert, R. Elschner, O. Stiewe, IH Baek et al.
C. Schmidt-Langhorst, J. Pfeiffer, R. Elschner, R. Emmerich, F. Chowanek, IH Baek, R. F. H. Fischer, C. Schubert
05 — Expertise
