Vlad-Marius G.

Data Scientist

660 dollar
7 years

My experience


Petnica Science CenterAugust 2014 - Present

• This was a two weeks research-based summer camp where students were assigned small projects related to the research interests of the different departments within the institution
• My group was concerned with the computational simulation of light quanta (photons) as interacting with different materials (a problem intensively studied in astrophysics)
• My role was to develop a Monte Carlo simulation in C concerned on the interaction between photons coming from distant objects and the Earth’s atmosphere
• The result was a graphical simulation of the Moon's halo effect.

AccessPayMay 2019 - Present

Building an Automated Cashflow Forecasting Solution aimed at improving the Cash Management Operations of large Enterprises.
Day to day activities:
- Building Data Collection, Data Cleaning and Data Analytics pipelines to leverage the data from the company’s legacy system
- Researching, Developing and Implementing Time Series Forecasting methodologies
- Managing Project: chairing team meetings, managing budgets and project delivery targets, raising client engagement, etc.
Key Skills: Python 3, PostgreSQL via AWS RDS, Agile PM

The University of ManchesterSeptember 2018 - January 2019

I did my 1st-semester MPhys research project on the topic of neutrino physics, working on the development of the Deep Underground Neutrino Experiment (DUNE).
The research directions included:
- making use of Monte Carlo techniques to simulate the liquid argon (LAr) detector environment
- running tests to check the suitability of the simulated geometries
- devise algorithms for a triggering system based on scintillation light yielding from neutrino interactions with LAr media
- prepare weekly presentations to showcase preliminary results
- write a thesis and defend it in an hour-long interview

Key Skills: Python 3, C++, Linux Operating System, Data Handling and Visualisation, Data Mining.

TU Dortmund UniversityJune 2018 - August 2018

My project was aimed at evaluating the potential of the D* (D0 (KK) π(slow))D(Kππ) final state for the CP violation measurement in B^0→D*D analysis using Run I data (2011-2012) of the LHCb detector (CERN, Geneva).

The analysis procedure included:
- Data cleansing to remove outliers and unphysical background;
- Training ML algorithms (Boosted Decision Trees) to find the best compromise between background reduction and signal efficiency, based on Monte Carlo simulations;
- Fitting data to extract the candidate B^0 events.

My results showed that the additional final state did not have enough statistics in the Run I data to be used for computing the CP violation parameter.
My main contribution was to develop and implement an algorithm for finding the best features to be used by the Boosted Decision Tree. This algorithm verified the goodness of some established kinematic features (e.g. decay length of D meson) and suggested new "interesting" attributes (e.g. number of D candidates within a cone of the secondary vertex) out of a number of about 5000 feature candidates.

Key skills: C++, PyRoot, Machine Learning, Data Science, Big Data

The University of ManchesterSeptember 2017 - May 2018

Together with 6 other students with various STEM backgrounds, we joined the CanSat project aiming to develop and build a "can size satellite" designed for atmospheric measurements.
My main contribution was in programming the Arduino Nano microprocessor to execute required tasks (collect data, store it on an SD card and send it to the ground station through an XBee radio). I have also contributed to assembling the components on a stripboard and soldering.
Our "CanSat" was tested in the UK CanSat Competition 2018, where it performed below expectations due to a weakly soldered connection. Our consolation was winning the prize for the best presentation within the competition.

The University of ManchesterOctober 2017 - January 2018

My role in this position is to welcome
the applicants for undergraduate studies to the School of Physics and
Astronomy and give them a brief introduction into the student life and
the opportunities for development within our university.

Self-employedAugust 2017 - September 2017

•     Tutored students in Mathematics and Physics both at high school level and undergraduate degree level.

Romanian Institute of Science and TechnologyJune 2017 - July 2017

I have joined Prof. Marius F. Danca in his research in the field of Nonlinear Science. During our collaboration, my contribution was acknowledged in two research articles:

• Marius-F. Danca, Nikolay Kuznetsov, “Hidden chaotic sets in a Hopfield neural system”, Chaos, Solitons and Fractals, 1st July 2017
- In this paper the authors were investigating a simplified Hopfield neural network (with three neurons), aiming to prove existence of hidden chaotic sets in the system;
- My role was to find a computationally feasible way to represent the basin of attraction of the system in the vicinity of the initial value of the chaotic hidden transient (1.9;3;1). With several days to complete a task that would take weeks on a classical computer, I realized I could modify the algorithm to run it on multiple computers. This action increased the speed of the computation exponentially, making the task achievable within a daytime.
-The attraction basin unveiled self-similarity (see figure), verifying a theoretical prediction given in the literature

• Danca et al., ‘Fractional-order PWC systems without zero Lyapunov exponents’, Nonlinear Dyn (2018) 92: 1061
- I have contributed to this research paper by carrying out some specific computational tasks, such as localizing the zeros of the Lyapunov exponents in terms of the input parameters of the dynamical systems that were analyzed.

Online Course - Data Science EssentialsJanuary 2017 - January 2017

Introductory course in Data Science and Predictive Analytics including:
•    Simulation and Hypothesis testing
• Data mining and Data Visualisation
• Data cleansing and manipulation
• Machine Learning techniques using Azure ML

My stack

Big Data

Big Data


Python 3.5, Matlab, C/C++

Analysis methods and tools



AWS, Machine Learning




Research, Data analysis, Data Science

My education and trainings

- - Licenses & Certifications

- - Volunteering

- - Courses

Master's degree (MPhys) - Physics Grade First (80) - The University of Manchester2015 - 2019

Online Course - Machine Learning - Coursera2017 - 2017

High School - Mathematics and Computer Science - International Computer High School of Bucharest2013 - 2015

- - Honors & Awards2015 - 2015