QuantStack

Matthias Meschede

Chief Operating Officer

Contributes to the data ecosystem in Python/Rust/Haskell/Nix

Isabel Paredes

Scientific Software Developer

Contributes to Jupyter/Emscripten-Forge/Robotics ecosystem

Founded in 2016 in Paris
Principal contributor to Jupyter, Apache Arrow, Mamba/Conda
Multi-stakeholder projects with millions of users
Open source development for clients
Other topics: C++, scientific computing, WebAssembly, interactive applications
Animation of PyData community

Empowering scientists and developers worldwide with cutting-edge tools

Accessible
No installation required - run Python, R, Julia, and 40+ languages directly in your web browser.
Literate programming
Combine code, rich text, equations, and visualizations in a single document that tells a story.
Interactive
Execute code cells individually, experiment iteratively, and see results immediately.
Visual
Create stunning plots, charts, and interactive widgets that bring data to life.
Global impact
Used by millions worldwide - from students learning to scientists discovering breakthrough insights.

🖥️

1. Full Application Framework

Jupyter has become a comprehensive stack for building advanced professional applications.

⚡

2. WebAssembly

Jupyter can entirely run client side, greatly simplifying deployments.

Serverless execution
Faster, easier to deploy, scaling infinitely, cheaper.
Enterprise deployments
No (complex) server, no maintenance, no scaling issues. No installation. Built-in security through browser sandboxing.
W3C Standard
WebAssembly is a W3C standard, stable and supported by all major browsers. It is not going away any time soon.
Most languages can compile to WebAssembly now but not all
Python, Rust, Fortran, R, C++, ...
Libraries start to become available but not all are there yet
See QuantStack's Emscripten-Forge for package management in the browser. Some libraries are not yet available and difficult to port.
Hardware not as easily accessible
Bluetooth, GPU and other hardware need to be accessed through specific interfaces (Web Bluetooth, WebGPU, etc.).
Extra hoops because of sandboxing
Accessing the file system, network, or other resources may require additional steps.
Limitation to user hardware
Server-based Jupyter can go beyond the user hardware, e.g. GPU, TPU, or HPC clusters.