Welcome.

I am Pierre Lagache, and I am passionate about creating simple yet elegant solutions to projects.
Scroll through to discover my journey and the stories behind each one.
Please feel free to contact me.

Machine Learning

System & network

Multilayer Perceptron

To achieve this project I wrote a machine learning framework composed of a “Tensor” class that perform automatic calculus of Derivatives.
The Idea is to leverage the chain rule to compute the gradient.
Finally, I will make an “Optimizer” class that will perform differents types of parameters optimization.

Data Science X Logistic Regression

The term Data Science in the title will be clearly considered by some to be extra. That is true. The subject is vast.
In DSLR, I trained a multi-classifier(one-vs-all) using a logistic regression.

  • I analyzed the dataset, displayed it in different ways, and selected and cleaned unnecessary information.

scatter-plot-example

  • I trained a logistic regression that solves classification problems. After that I created an interface to edit parameters for training.

gradio-interface

Inception of Things

Inception of Things is a system administration project designed to set up multiple environments according to specific rules. The entire codebase is executed within a virtual machine, ensuring isolated and controlled settings. I also delved into the different types of virtualization available, particularly libvirt, KVM and QEMU.

  • I built a virtualized environment with Vagrant and managed its different VMs with K3s.
  • I created Ingress rules to control how traffic is directed to different replicas.
  • I implemented Continuous Integration and Continuous Deployment using K3d and Argo CD. K3d allowed me to run a lightweight Kubernetes cluster in Docker, while Argo CD automated the deployment of applications

Ping

This project is about recoding the ping command with the C language.
I chosed to reimplement ping in C, as it allows me to control system resources and handle network operations manually.

  • I manipulated various internet protocols such has ipv4 and icmp.
  • I implemented the creating and sending of socket.
  • I developed the parsing for different icmp echo responses.
  • I implemented time management to determine the round-trip time.

ping_command_example

Linear Regression

I implemented a linear regression to model the relationship between variables. The model was trained using the gradient descent algorithm.

  • I created a second program that retrieves one variable based on the other.
  • I developed a program that generates a GIF of 120 images of the plots during training.

linear_gif

Contact and informations

Mobile: 06 87 06 82 82
Email: plagache@protonmail.com
Profile github: https://github.com/plagache
Profile malt: https://www.malt.fr/profile/pierrelagache1