ATLAS open data • Higgs (H → ZZ → 4ℓ) • ROOT • TensorFlow/Keras • Ares/PLGrid
I work at the intersection of experimental high‑energy physics and machine learning. Recent work includes discriminating Higgs events from Standard Model backgrounds in ATLAS 13 TeV open data using deep neural networks and ROOT‑based feature engineering.
Higgs, electroweak signatures, multilepton final states, significance estimation.
Supervised learning for S/B separation, ROC/AUC analysis, calibration and uncertainty.
ROOT, Python, TensorFlow/Keras, scikit‑learn, Hypatia; HPC on Ares (PLGrid).
Introductory quantum computing (Qiskit) and numerical methods (Quantum ESPRESSO basics).
Hands‑on work with ATLAS open data and comparative studies of shallow vs deep networks across HEP datasets.
Built a Keras pipeline to classify H → ZZ → 4ℓ events vs. SM backgrounds using features extracted with ROOT. Achieved AUC ≈ 0.94 and discovery significance ≈ 5.37σ. Produced ROC curves and score distributions showing clear separation.
Compared models on ATLAS (AUC ~0.88), MiniBooNE (~0.98), and Belle II (~0.987). Used Cyfronet Ares for accelerated training (epoch: CPU 150s → GPU 23s). Conclusion: deep NNs excel on complex, large datasets; shallow remain competitive when features are physics‑informed.
Implemented baselines (logistic regression, decision trees, DNNs) and profiled CPU/GPU scaling; emphasized reproducible evaluation and ROC/AUC as standard metrics.
Completed short workshops covering ab‑initio simulations and fundamentals of quantum circuits and algorithms.
Programming, data/HEP tools, systems, and writing.
Python, C, C++, HTML, MATLAB
ROOT, Hypatia, TensorFlow, Keras, scikit‑learn
Linux, Git/GitHub, Docker, SLURM
LaTeX
Interdisciplinary school: precision spectroscopy, quantum information, open systems, complexity.
160‑hour program with ROOT training, Hypatia hands‑on, and supervised ML project.
Lectures on ML, high‑dimensional inference, and generative‑AI perspectives.
Represented Nepal in theoretical physics competition at Trinity College Dublin.
Open to research internships and collaborations in collider physics and ML for HEP.