By Jakub on February 6, 2020
At Deepnote, we are building a data science platform where data scientists can collaborate in real time and solve the most difficult AI and ML problems. We believe the future of data science is collaborative, supported by powerful tools specifically designed for data exploration and model prototyping.
Right now, as data scientists, we have to rely on the tools inherited from the software engineering world to solve the problems we are working on. But the data science workflow is fundamentally different. We work with large datasets, run hundreds of experiments, use powerful GPUs. We work in teams, we share our findings with others. We need to collaborate.
The tools we use today were not built for this purpose. Even as the number of data scientists grows rapidly each year, there isn’t a tool which is collaborative at heart, takes care of the infrastructure and scales with you as the complexity of your data science projects grows. That’s why we’re building one. We are building a notebook that lets you focus on your work. We are building a notebook that makes you a better data scientist.
How we got here
As data scientists and software engineers, we started to notice the discrepancy between the level of tools available to software engineers and data scientists, but also between the quality of data science tooling available to employees of the giants like Google, Amazon, Facebook or Netflix, and between what is available to everyone else. It’s not just better resources what’s keeping these companies ahead — it’s better tooling, too.
We started Deepnote at the beginning of 2019. Building on top of the great ideas behind Jupyter, we put together our own experiences from companies like Two Sigma, Palantir, Mozilla, Skyscanner, or Google, to bring the state-of-the-art tools to every data scientist in the world.
Deepnote is an enhanced, collaborative and Jupyter-compatible cloud-based notebook. Now, one year later, thousands of users and the biggest names in the data science community are already using Deepnote.
Peter Norvig, Director of Research at Google is using Deepnote to give wide audience access to his personal projects on his github page. And he’s not the only one — in fact, if you go to pretty much any popular repo with Jupyter Notebooks on Github, chances are you will find links to Deepnote in there. Students use Deepnote for their coursework, authors use Deepnote to publish interactive exercises from books on machine learning. Open source developers use Deepnote to publish their computer vision libraries online.
Data scientists love to share their work with others. We’re making it easy for them.
We’re excited to partner with Accel and Index as the lead investors of our seed round, along with Y Combinator and Credo Ventures. In addition, we are joined by some of our friends Greg Brockman (OpenAI), Dylan Field (Figma), Elad Gil, Naval Ravikant, Daniel Gross, Lachy Groom and others.
What lies ahead
This is the beginning.
We understand the needs of a data scientist or a machine learning expert are different to those of a software engineer, and we want to bring the ideas found in the best software engineering tooling to data science while preserving the unique workflow that data science has.
A data science project is not just the code itself, it’s the combination of code, data and execution environment. Notebooks represent a new kind of computational platform. An interactive document, powerful enough to solve the most difficult problems and collaborative enough to share your findings with others. The best data science notebook would allow you to reproduce your findings, make it easy for you to version your models and make you a better data scientist by allowing you to focus on what matters.
This is the notebook we are building.
Join us if you want to be part of this story.
Share this post