Anthony Bwembya

Experimental Physicist · Scientific Computing · Signal Processing · Instrumentation · AI & Data

Education

Doctoral Research in High-Energy Physics

Radboud University, Netherlands
2021 – 2026

MSc, Nuclear Physics and Technology

National Research Nuclear University MEPhI, Russia
2018 – 2021

BSc, Physics

University of Zambia
2013 – 2017

Languages

  • English (Native · preferred and international working language)
  • Russian (Professional working level)
  • Dutch (Developing)
  • Bemba (Native)
  • Nyanja (Native)
  • Lozi (Native)

Focus Areas

  • Intelligent sensor systems
  • Radio interferometry
  • Scientific machine learning
  • Sustainable technology
  • Nuclear technology
  • Detector development

Projects

A selection of technical work spanning sensor calibration, scientific modelling, AI-assisted diagnostics, and sustainability.

Abstract artwork inspired by timing stations and radio signals Abstract artwork inspired by a distributed antenna array

Visual studies inspired by distributed timing, antennas, and coherent signals.

Radio Interferometry for Cosmic-Ray Detection

Coherent reconstruction of inclined air showers using synchronised measurements from widely separated antennas.

SENSING

Production research workflow.

Precision Timing and Phase Calibration

Beacon- and broadcast-carrier methods, discrete phase solutions, and closure diagnostics for distributed detector stations.

SENSING

Validated research workflow.

Antenna Response and Array-Systematics Studies

Quantified how antenna models, geometry, refractivity, noise, and calibration choices affect reconstructed signals.

MODELLING

Cosmic-Ray Propagation Modelling

Investigated the positron excess using PAMELA/AMS-02 observations and GALPROP simulations.

MODELLING

Water-Level Monitoring for Energy Security

Awarded a £5,000 grant for satellite-imaging and geodesy methods supporting water-level tracking in Zambia.

IMPACT

AI-Assisted RFI Classification

Concept development for probabilistic interference classification and RFI-aware calibration in MeerKAT/SKA-relevant radio data.

AI

Research concept in development - not yet deployed.

Machine-Learning Case Study

Problem

Radio-frequency interference can corrupt calibration and astronomical measurements, while simple threshold flagging discards useful data.

Approach

Concept design using spectral, temporal and modulation features with supervised classification, calibrated probabilities and anomaly detection.

Evaluation plan

Benchmark precision, recall and calibration residuals against established flagging methods, then test probabilistic weighting in the calibration pipeline.

Maturity

Research concept and prototype plan. No deployment or performance metric is claimed yet.

Working Principles

Traceable

Explicit assumptions, versioned inputs, reproducible pipelines, and clear provenance from raw data to result.

Uncertainty-aware

Calibration errors, model limitations, and confidence are treated as part of the result—not as afterthoughts.

Operational

Diagnostics and visualisations are designed to help people detect failure modes and make better decisions.

Responsible

Models should be explainable, appropriately validated, and used with attention to scientific and societal consequences.