Simon Bührer
MSc student at ETH Zürich working on differentiable logic gate networks, sequence models, and learning systems that run on tight hardware budgets.
I study electrical engineering and information technology at ETH Zürich, with research at the intersection of machine learning and hardware. My recent work centres on differentiable Boolean computation: training neural networks whose primitives are logic gates and lookup tables rather than continuous MACs, so that the resulting models map cleanly to FPGAs and other low-energy substrates.
Outside of that, I have written about transformer architectures for multivariate time-series classification, the role of replay sampling in catastrophic forgetting, and automata-inspired RNNs that can perform exact arithmetic.
Research interests
- Differentiable logic gate networks and FPGA-native neural architectures.
- Recurrent and transformer-based sequence models, including automata-inspired RNNs for exact computation.
- Continual learning and replay-based mitigation of catastrophic forgetting.
Experience
- 2025 – 2026
Everllence
Internship → Werkstudent · Zürich
Seven-month industry internship followed by a three-month part-time continuation.
- 2022 – 2023
Fokusprojekt · Dübendorf
Bi-liquid rocket engine: design, instrumentation, throttle control. 677 N thrust, reproducible firings.
- 2021 – 2022
ETH Zürich
Teaching Assistant — Engineering Mechanics
- 2020
Bühler Group
Internship · London, UK
- 2019
Swiss Armed Forces
Panzerjägersoldat, Aufklärungskompanie · Thun
Education
- 2024 – 2026
ETH Zürich — MSc Electrical Engineering & Information Technology
- 2020 – 2023
ETH Zürich — BSc Electrical Engineering & Information Technology
- 2015 – 2019
Kantonsschule Kreuzlingen — Gymnasiale Maturität
Theoretical and Mathematical Physics
Languages
German (native) · English (professional) · French (limited) · Russian (elementary)