Bijan Berenji, PhD
Bijan.berenji@fizisim.com www.fizisim.com
206.537.5509
Career: Research scientist, research consultant, and information technology consultant
Education and training: Applied physics PhD – Stanford University, 2011 – nuclear physics, quantum gravity, astronomy bijanb@alumni.stanford.edu
- Bachelor of Science (BS) physics, UCLA, 2001
- Radiological sciences postdoc– UC Irvine, 2012
- Nuclear astrophysics PhD – Stanford University, 2011
- Applied mathematics Postdoc – UCLA, 2013
- Community college professor, El Camino College, Torrance, CA 2014
- Adjunct physics Professor, California State University, Los Angeles, 2015-2018
- Institute for Nuclear Theory visiting fellow – University of Washington, Seattle 2020
- Machine learning / data science lead and mathematical modeling at fiziSim LLC 2016-current
Professional experience and competencies: Information technology – machine learning – data science — research consulting — computational physics – theoretical nuclear astrophysics – quantum computing — medical imaging — applied mathematics and applied statistics– data analysis – data analytics – scientific programming — experimental physics — management – university professor — online/remote teaching – social science modeling – electronics and electromagnetics
CTO and CEO for startup firm fiziSim LLC – www.fiziSim.com – incorporated in California and state of Washington – for hire on corp-to-corp basis
Technical Experience and projects (short form):
- Website design – technical support – medical billing solutions
- Python engineer – machine learning engineer — scientific computing expert– object-oriented programming – data science – artificial intelligence engineer
- Python – C++ — Matlab – Mathematica — perl – Linux 500,00+ lines of code
- Monte Carlo simulation algorithms
- Numerical python – numpy – matplotlib – scipy – jupyter — github
- C++ compiled with gcc (gnu compiler) – visual studio .NET
- Matlab imaging routines – differential equations – image analysis routines
- Linux – Ubuntu 20.04 – Scientific Linux – Red Hat Enterprise Linux (RHEL)
- Windows 10 – DOS — MacOsr
- Former professor for physics computing and online/remote teaching, introductory and advanced physics, at California State University, Los Angeles www.calstatelaphysics.org www.profberenjicsulaphysics.com
- Python / C++ / matlab / jupyter / github / perl / ubuntu /
- theoretical physics algorithms – neutron star modeling – gravitation – NASA and Department of Energy Projects – Fermi Large Area Telescope and LIGO
- Machine learning in gamma-ray and electron energy classification – experimental physics
- computational medical imaging – positron emission tomography (PET), computed tomography (CT) imaging, X-ray
- artificial intelligence: neural networks, parameter estimation
- nuclear physics laboratory – data acquisition and electronics, pipeline
Projects
Data science: For Stanford University as my client, I wrote data analysis pipeline in python to sort through PB dataset live from an astronomical observatory at NASA;I linked python code against C++ scientific libraries using SCons and Cmake. ; I analyzed the data with data analytics in python using numpy as matplotlib; advanced statistical modeling using Poisson statistics, Bayesian inference, and log-likelihood statistical methods; Monte carlo simulation of PB of simulated data using C++ linked against the nuclear physics package ROOT
responsibilities: coding, mathematical modeling, machine learning
Machine learning: For Stanford University as my client, I wrote code to reconstruct the estimatated value of an input parameter (photon energy) given over 200+ variables in a C++-based statistical and machine-learning framework of a nuclear physics experiment. The objective was to reconstruct the best estimate of a gamma-ray photon’s energy in an experimental situation where the absolute truth value is not known 100% correctly. Various input variables were fed into a statistical model, and the estimated energy was output. The model was refined to minimize the statistical deviation between the Monte Carlo known energy value and the reconstructed parameter.
Responsibilities: coding, machine learning
Mathematical modeling: For fiziSim LLC and University of Washington, Seattle as clients, a relativistic and strong-force QCD-model of theoretical physics was used to estimate the detectabilitity and mass of a theoretical particle known as the axion. The mathematical model was used to generate Monte Carlo-based datasets which were then analyzed in python.
Responsibilities: mathematical modeling
Medical imaging:
Medical nuclear physics: For UCLA, mathematical model of radioactive sugar molecules was used to image cancer cells
responsibilities: coding, biochemical modeling, chemical physics
Medical CT scanning: For UC Irvine, using attenuation coefficients, estimate concentrations of biological molecules using a machine learning and mathematical modeling from CT image data.
Responsibilies: coding, medical imaging projects
Teaching
I was a professor of C++, python, and Matlab for computing in physics courses that I taught at California State University. Topics included machine learning and numerical methods for undergraduate and graduate courses.
Responsibilities: coding, machine learning, physics computing, teaching, mentoring