MSc Engineering Physics @ Uppsala University: computational science and machine learning applied to physics and engineering.
- Data scientist (part-time) @ Locus Energy
- Engineering intern @ Redpine
- ML research assistant @ Uppsala University
- Bachelor's thesis @ ATLAS UU Group/ CERN
Data scientist (part-time): Locus Energy, Stockholm Building probabilistic forecasting models (Bayesian methods, state-space models, LSTMs) to predict revenue from intermittent renewable generation under market volatility.
ML research assistant: Department of Information Technology, Uppsala University Developing deep learning models that classify cardiac arrhythmias from 3M+ ECG recordings, mapping rhythm disorders to their electrical signal patterns to support clinical decisions.
Bachelor's thesis: ATLAS Particle Physics Group, Uppsala University (paper)
Developed neuromorphic learning models to detect and reconstruct particle signals from data collected at CERN's Large Hadron Collider. Benchmarked neural network architectures on large-scale, sparse, high-multiplicity physics datasets.
Particle track reconstruction in LHC detector data using a two-stage filtering neural network
Spiking neural network neurons learning to fire: membrane potentials and spike events over time
ML research assistant — Uppsala University
U-Net segmentation model that extracts ECG waveform traces from plot images, trained on the PTB-XL dataset — a step toward automatically digitising ECG rhythm data for arrhythmia classification.
Ground-truth trace mask (top) and the trained model's prediction overlaid on the original ECG plot (bottom, purple = predicted and true trace agree)
Multi-agent disinformation tracker — Top 5, Google × Epiminds Hackathon A multi-agent retrieval system using Claude to scrape and analyse web sources in real time for disinformation detection.
VSCode developer assistant — 2nd place, KTH AI Society × Florent Venture Hackathon A VSCode extension with a React frontend and a Claude-powered backend for real-time developer assistance.
Open-source eye tracker — Winner (local round), Antler's National Start-up Buildathon Built in React, capturing and storing gaze-coordinate data for cognitive analysis.
A physics and mathematics foundation with a computational science and AI specialisation — probabilistic ML, numerical methods, and scientific computing.
Numerical solution of the 2D acoustic wave equation using SBP-projection methods
Course projects, research code, and hackathon submissions. Mostly Python, kept readable and reproducible.
- Quantitative & ML: Deep Learning · Time-Series Forecasting · Signal Processing · Optimization & Monte Carlo Methods
- Programming & Data: Python (NumPy/Pandas, PyTorch, scikit-learn) · SQL · Git (CI/CD) · Linux/Bash · HPC (SLURM)