(CG_dmitriy/Shutterstock)
Marimo has emerged from stealth with $5 million in seed funding. The startup has launched an open-source Python pocket book designed to offer builders with a instrument that’s reproducible and Git-friendly. It may also be deployed for manufacturing as an internet app and executed as a script.
The seed funding spherical was led by Anthony Goldbloom (ex-Kaggle, Sumble) and Shyam Mani of AIX Ventures, with participation from Jeff Dean (Google), Clement Delangue (HuggingFace), and different notable traders.
Python has been the dominant language in AI. In October 2024, it overtook JavaScript as probably the most used language in GitHub. Because the lingua franca of AI, machine studying (ML), and STEM, Python’s adoption has continued to broaden as these applied sciences mature and transition into manufacturing environments.
To help this transition, a number of venture-backed startups are modernizing key elements of the Python stack. MotherDuck and Polars are constructing quick OLAP engines to exchange Pandas. Astral and Prefix are modernizing package deal administration by changing legacy installers like pip with UV and Pixi. Modal and Coiled are additionally working to allow serverless capabilities to be used at any scale.
Whereas there was vital modernization of Python instruments to help AI, ML, and information science, the Python pocket book, which serves as the first surroundings for a lot of this work, has acquired little consideration.
(Iurii-Motov/Shutterstock)
Marimo plans to repair this with its Python pocket book addressing the important thing shortcomings of conventional notebooks like Jupyter. Marimo claims that, not like Jupyter, the marimo pocket book ensures true reproducibility by retaining monitor of dependencies and execution circulation routinely. This might help remove frequent problems with Python notebooks akin to damaged code from lacking dependencies, difficulties with model management, and the dearth of reproducibility.
In response to Marimo, their pocket book not solely addresses the important thing limitations of conventional instruments like Jupyter but in addition outperforms platforms akin to Streamlit and Gradio. The marimo pocket book comes with an AI assistant conscious of dataframe schemas and integration with industrial coding copilots like GitHub Copilot and Codeium.
Moreover, marimo goals to offer a extra user-friendly and expressive surroundings all through the lifecycle, from deployment to manufacturing. It additionally comes with a big library of UI widgets like information body transformers and cheat parts. This strategy addresses the restrictions of static and error-prone notebooks by enhancing interactivity.
“After I began engaged on marimo, I had simply completed a PhD at Stanford the place I helped develop an open-source library for optimization to one million month-to-month downloads, and earlier than that labored on TensorFlow at Google Mind,” shared Akshay Agrawal, Founder and CEO, Marimo Inc., by way of a weblog publish on the seed announcement. “Jupyter notebooks had been important to my work, as a result of they let me see my information whereas I labored on it — however they had been additionally extraordinarily irritating.”
“Hidden state led to hidden bugs (over a 3rd of the ten million Jupyter notebooks on GitHub fail to breed), the JSON file format made them exhausting to make use of in Python codebases, and the ultimate paperwork lacked interactivity. Whereas Jupyter notebooks are extensively used for AI/ML improvement, STEM, and information engineering, now greater than ever, there’s rising consensus that this type of work shouldn’t be accomplished in error-prone scratchpads.”
(BEST-BACKGROUNDS/Shutterstock)
Akshay, together with co-founder Myles Scolnick, wished to create a programming surroundings that “blended one of the best elements of interactive computing with the rigor of conventional software program improvement.”
Inside a yr of its launch, Marimo has been downloaded over 1 million instances. It’s getting used at a number of organizations and universities across the globe, together with BlackRock, Mozilla AI, MLB, and Stanford. The web group can also be rising robust, with over 7.8k stars on Github, 60+ contributors, and 1k members on Discord.
Marimo plans to make use of the funds to proceed growing the Marimo pocket book and discover creating different information instruments. As the corporate focuses on refining the pocket book, they do have plans to launch a paid product sooner or later. Nevertheless, the corporate stays dedicated to open-source improvement, making certain its instruments are free and accessible.
Associated Objects
Anaconda’s New Device Lets Customers Run Python Code Inside Excel
Python Now a First-Class Language on Spark, Databricks Says
KX Goals to Unlock Python’s Potential with PyKX 3.0