Academic Staff

Supraja Ramesh

Karlsruhe Institute of Technology (KIT)

Institute of Telematics / TECO

Vincenz-Prießnitz-Straße 1

76131 Karlsruhe, Germany

Building 07.07

Room 202

rameshδteco.edu

Profile

I am a PhD student working on earable technology for health applications, with a focus on biosignal processing and machine learning methods for robust, real-world health monitoring. Feel free to contact me if you’re looking for thesis topics in earables, machine learning, or biosignal analysis.

Short CV

since 2024 PhD Candidate at TECO
2022 – 2024 M.Sc. Medical Technology at Friedrich-Alexander-University Erlangen-Nuremberg
2018 – 2022 (Senior) Software Engineer at Freshworks (India)
2014 – 2018 B.Tech. Electrical & Electronics Engineering at SASTRA Deemed University (India)

Projects

since 2024 KARE
since 2024 Multi-modal stutter detection
since 2025 Respiration rate detection across sensor modalities

Research Interests

Biosignals
Deep Learning
Wearables

Teaching

Proseminar: Mobile Computing (WS/SS)

Proseminar: Designing and Conducting Experimental Studies (WS/SS)

Completed / Ongoing Theses

MA: Target Speaker Separation on edge devices (Elham Khosravi, 2024/25)

BA: Key Influencers of Consistent Physical Activity engagement (Jialu Cheng, 2025/26)

BA: Benchmarking TinyML models for PPG-based blood glucose level estimation (Markus Neufeld, 2025/26)

BA: Multimodal ML analyses to distinguish between nasal and mouth breathing with OpenEarable 2.0 (Aloïs Dieval-Lozac'h, 2025/26)

BA: Closed-Loop Rhythmic Haptic Biofeedback via Smartwatch for Sleep Onset (Iskander Rakhimov, 2026)

Theses

Unpacking PPG Signals for Transparent Blood Glucose Level Sensing.

OpenBachelor

Diabetes is a major global health challenge that requires careful blood glucose monitoring alongside lifestyle interventions. Photoplethysmography (PPG) offers a promising, non-invasive way to estimate blood glucose levels using machine learning in wearables. Current PPG-based methods mostly rely on complex black-box models that obscure which signal features are actually important. This thesis aims to identify and analyze key PPG features that reliably contribute to blood glucose level estimation.

Topic Areas

Biosignal Analysis and TechnologiesData Mining

Publications

2025

Closed-Loop Rhythmic Haptic Biofeedback via Smartwatch for Relaxation and Sleep Onset
Lee, J.; Moschina, D.; Ramesh, S.; Röddiger, T.; Kunze, K.; Beigl, M.
2025. Proceedings of the ACM International Symposium on Wearable Computers (ISWC’25), 30–37, Association for Computing Machinery (ACM). doi:10.1145/3715071.3750412