About me
pwd
/home/artorithme/myWebsite/
whoami Jean-Pierre Axel Gninahophin
cat activities.md I spend most of my time messing with my computer when I’m not biking.
chmod +x presentationScript.sh ./presentationScript.sh WELCOME TO MY BLOG !
You’re currently on the about me page!
Click on the Artorithme logo to explore my blog!
What I’m good at
Programming
This is what I do during my weekends
When I want to have fun
If I’m a not programming, I’m probably reading/writing about it.
Analysing Data and explaining Models
I really enjoy exploring data, and sharing insights I gain from it.
I like building fancy models, using Neural networks and Xgboost, but what I like the most is making them understandable to drive better decisions.
Education
2020
PCSI-PSI
Introductory class in Mathematics, Physics,Engineering
- Analysis, Algebra, Geometry, and Probability
- Mechanics,Thermodynamics,Electricity, Magnetism,Fluid Mechanics and Optics
- Computer Science (Python) and Mechanics
2022
Actuarial science - DUAS
Advanced classes in Probability Theory, Statistical Analysis, and Risk Modelling
Introductory classes in Finance,Insurance and Investments
Machine Learning Engineer Certification
Codecademy
2024
Master in Signal and Image Processing : Specialization in Computer-vision
Programming languages I can’t leave without
R
R is my favorite language by far and the main reason I created this blog. I think people don’t realise how fantastic it is because they discover it in Academia with a statistic teacher who isn’t really interested in the programming aspect.
Javascript
Javascript is a mess. Everyday,there’s a new UI library or a framework… Npm will have you dowload the whole internet before you can center a div. But that’s the price to pay to have nice things on the internet.
Python
Python doesn’t really need any presentation.
Bash
I’m a Gnu/Linux user and no, I don’t use Arch (by the way) because I have matured and I use Debian as all my sysadmins friends. For me, Bash is a lifestyle.
A 2D array of my latest endeavors
| Title | Tech | Tech stack | Details |
|---|---|---|---|
| My personal website | HTML, CSS, Javascript,Bash,Python | Jquery, Bootsrap,Pandoc,pandas | |
| Monte-carlo Simulator | C++ | STD | A Monte Carlo simulator in C++ designed to estimate the expected loss or gain from investing in a security whose value varies according to a leptokurtic distribution. |
| Advanced social media Web-scrapper | Python, Javascript, Chromium | Selenium, Pandas | A web scraper that mimics typical user behavior to gather data and assist in combating scam artists. |
| Reverse engineering an api to scrape chess games and train a Chess-bot | Bash, Php | jq, curl | |
| Insurance pricing dashboard | HTML, CSS, Javascript, R | shiny, data.table, Bootsrap | I significantly optimized the performance of a complex pricing tool by migrating it to R, resulting in real-time update capabilities, and developed an intuitive R dashboard to simplify user interaction and enhance overall all within a remarkably short timeframe. |
| Recovered a non-booting computer by resolving a kernel failure and repurposed it as a functional Linux backup server. | Bash, OpenSSH | ||
| Developed and optimized a multi-layer neural network model to predict wine quality, utilizing the R port of PyTorch, and the H2O library | R, Python | Torch, Pytorch, R6, H2O, data.table | Trained a multi-layer neural network on a severely imbalanced dataset, leveraging insights from relevant literature to develop a high-bias, low-variance model that effectively distinguished high-quality wines from the rest, despite the datasets inherent class imbalance. |
| Designed a modular framework leveraging R6 object-oriented programming and grid search to programmatically generate and evaluate multiple model configurations, enabling the identification of the optimal model architecture. | R, Python | Torch, Pytorch, R6, H2O, data.table | To support my research, I created a tool to programmatically generate neural network models in Torch, which enabled me to systematically explore different architectures and identify the optimal configuration. Through a grid search on a subspace of multi-layer neural networks, I was able to find a model that effectively balanced complexity, performance, and prediction time. |
| Conducted an in-depth analysis of NFC technology | NFC, C, Bash | Delved into the inner workings of NFC technology, examining its fundamental physics principles and circuit design to understand its behavior. Furthermore, conducted a security-focused technological watch to evaluate the resilience of NFC-enabled devices against potential threats, simulating attack scenarios and identifying areas of vulnerability that could be targeted by attackers. | |
| Conducted a data-driven analysis to investigate the relationship between smoking habits and social category in an insurance portfolio, challenging common heuristics that could lead to unfair pricing practices. | Python, R, scikit-learn, data.table, matplotlib, seaborn | ||
| Developed a photoplethysmography algorithm utilizing computer-vision techniques to extract blood frequency signals from smartphone video recordings | Python, OpenCV | ||
| Designed and implemented data engineering solutions for financial data related to the European Union Emissions Trading System (EU ETS) | Python, R | data.table, pandas, scrappy, requests | Conducted a comprehensive review of existing literature on the European Union Emissions Trading System (EU-ETS), and collected, processed, and analyzed relevant data to support further research and analysis during an R&D internship. |
| Image Compression using Singular Value Decomposition (SVD) | Python | numpy |