Hi I'm Manu Potrel
I'm an ML Engineer
and a teacher
My favorite tech topics include Deep Learning, Large Language Models/Large Multimodel Models and Software Engineering
Get in touch →
Want to know more?
A bit about me.
I am . . .
01
An MLOps engineer
I have experience in all of the ML lifecycle, from designing a model to running and monitoring experiments and production inference. I have used all the classic frameworks and libraries in Python, and also some lesser known or less ML-centered ones, as discussed in the Tech section.
02
A DL engineer
I have worked a lot with Deep Neural Networks, especially in Natural Language Processing, though I also have some experience in Computer Vision. I have also had the opportunity to work on Large Language Models.
03
A self taught developer
I learnt how to code when I decided to switch from my position as a high school and university math teacher to the very promising field of Artificial Intelligence.
Python has been the de facto language for Machine Learning for a few years now. This has led me to practice it a lot and fall in love with programming in general. I can design complex projects and review dense merge/pull requests in Python, ML related or not.
I have collected a bunch of CS certifications that you can check out on my LinkedIn profile.
04
A software engineering enthousiast
I am very curious about everything related to software engineering. I entered this world with pure ML research in mind at first, but developed a passion for clean code, system design and workflow improvement. I always have one or two ongoing projects, hands-on or educational, and several others patiently waiting for their turn. While I firmly believe in specializing in specific areas, I also want a decent general knowledge of everything.
05
A passionate teacher
It started very early in my life. At 15 I was a teacher for my chess club and I managed to keep it going until I was 23. During my studies I was also a math private tutor. My first work experience was that of an actual math teacher (agrégé de mathématiques) in high school and university. Last but not least, as a second job, I am also working for a bootcamp company as a project mentor.
Resume
Relevant work experiences
I have worked as a . . .
Data Science developer
- (2021 - ) Sancare
I have been a Data Science developer at Sancare, a french company working with hospitals to help them categorize patient stays using NLP and structured data, and with pharmaceutical companies to run statistical studies.
- I have helped develop and design an LLM repository for entity extractions in pharmaceutical studies
- I have built components of the ML stack, reviewed critical packages and I have been involved in refactoring several key parts of the main ML packages
- I have run experiments to improve the performance of our NLP models
- I have monitored, debugged and maintained our ML services on the hospital prod servers
ML project mentor
- (2023 - ) DataScientest
This was the bootcamp company I chose to learn about Data Science, so it meant a lot to me when they reached out to ask if I would be willing to mentor new learners. Obviously, I immediately accepted.
I have been mentoring several projects based on story telling, data analysis, feature engineering and Deep Learning.
One of the most wonderful aspects of this job is that I have had many groups so far from different countries, so this has been (and keeps being) an incredible experience!
Math teacher
- (2017 - 2018) Lycée François 1er
- (2018 - 2021) Lycée Galilée
- (2019 - 2021) CYU university
I have given math lessons in university and engineering school, and I have been a match teacher in high school for over four years, where I had to learn how to communicate with a sometimes difficult audience of heterogenous levels (believe it or not, teaching those sometimes difficult teenagers has actually been one of the most enjoyable time of my career). I have also given some basic high school computer science courses during my time as a teacher.
Technical toolbox
Languages and frameworks knowledge
I have experience with . . .
Programming Languages
- Proficient: Python
- Decent: Rust
- WIP: Javascript, Typescript, C
- Coming soon: C++, Golang, Haskell, Lua
ML
- Machine Learning: pandas/polars, numpy, seaborn, scikit-learn, mlflow
- ML Ops: Monitoring, evaluating, fine tuning, investigating
- Deep Learning: tensorflow, torch, skorch, transformers
- LLM: transformers, langchain, rellm
Code Health
- Books read: Clean Code, Head First Design Patterns
- Certifications: The Software Designer Mindset
- Coming soon: The Software Architect Mindset, Grokking Concurrency
Workflow
- Code editors: NeoVim, Helix, VS Code, PyCharm
- OS: Ubuntu, MacOs, ArchLinux, Windows
- Shell & Terminal: Kitty/Alacritty, Zsh, Tmux
Web development technologies
- Known languages: HTML, CSS, Javascript
- Known frameworks/libraries: React (barely), Svelte (OK)
- Used once or twice: PHP, Nginx
Languages
- Native: French
- Professional: English
- High school (so please don't): German
A few of my personal projects.
Curious to see my projects?
My github page
This website
This website was created using Svelte with Tailwind CSS. I wanted to learn about some web technologies, being already somewhat familiar with HTML, CSS and JavaScript, so I thought "why not use one of the current cool kids' framwork?".
Click to check it out!
Management app
(Coming soon) I had been trying to learn Rust for a few weeks when I decided to follow ThePrimeagen's advice to start by building something I need. And while I do love Obsidian, I thought building a custom task management and note taking app would be both challenging and instructive.
Click to check it out!
Dotfiles
My most important config files.
Workflow improvement has basically become a hobby. I am primarily working on an Ubuntu OS, MacOS being a close second, and occasionally Arch as well.
Most of my workflow is keyboard and terminal based.
Click to check it out!
When work is over
My hobbies
I love spending time . . .