Razu Mohiuddin

Project 8 Collaboration

The Project 8 experiment aims to measure the absolute mass of the neutrino through the beta decay of tritium. It uses a novel technique called cyclotron radiation emission spectroscopy (CRES) to measure the energy spectrum of emitted electrons with unprecedented precision.

Within the collaboration, I led machine learning efforts including a U-Net-based deep convolutional autoencoder for segmenting class-imbalanced spectrogram images, improving efficiency over traditional methods by over 20%. I pioneered novel applications of NLP techniques for complex pattern detection in noisy signals, and explored physics-informed generative AI for reproducing electromagnetic signals from particle interactions. I also contributed to simulation projects including resonant cavity mode filtering and sensitivity analysis.

Recent publications are listed in the Publication section.

Roche Diabetes Care Internship

In Summer 2024, I interned at Roche Diabetes Care Inc. in the Mathematics, Algorithms & Data Sciences department, focusing on algorithms and advanced analytics.

• Developed virtual patient models from real-world clinical data
• Built predictive models for diabetes management tools
• Assessed novel techniques to reduce external dependencies
• Performed independent data analysis and model building
• Reported findings to the project team, department, and R&D

Previously, I also worked with the Gravitational Wave International Committee (GWIC) and the Community of Physics.