Susa Ventures is excited to announce our investment in Modelbit, a San Francisco-based startup building products that make it easy for data scientists to deploy and manage their models in production. We led the seed round, with participation from great funds like Sunflower Capital, Kearny Jackson, Weekend Fund, and Homebrew. We’re thrilled to (re-)welcome Harry, Tom, and the rest of the Modelbit team to the Susa Family.

Modelbit Co-founders: Tom O’Neill and Harry Glaser

This investment is an especially sentimental one for me because Harry and Tom’s previous startup, Periscope Data, was actually Susa Ventures’ first investment ever back in early 2013. Harry and Tom did an amazing job of building Periscope, and sold it for almost $150m in 2019. Despite having a big success under their belts, they were eager to build their next startup and to have an even bigger impact.

Their new company, Modelbit, makes it dead simple for data scientists to push their models to production without depending on ML Engineers. We at Susa believe that products like Modelbit will become increasingly critical as machine learning continues making a positive impact on every sector of the economy. The status quo at large tech companies like Uber is that data scientists build and train models, and then ML Engineers (MLEs) take over to productize and maintain those models. Hiring a team of expensive MLEs works if you have a big market cap and tons of money in the bank, but it won’t work for most smaller companies that want to deploy ML models, and that’s where better tooling for data scientists becomes incredibly important.

Modelbit’s product helps data scientists push their work to production without knowing anything about Docker, CI/CD, and the like. It’s so easy to use that even a VC like me who hasn’t written much code for almost a decade was able to get a model running in production in about 10 minutes.

“So simple even a VC could use it” is a testament to Modelbit’s great UX, but it’s just the starting point. Harry and Tom’s vision is to build a Heroku-like platform for data scientists: an easy to use, reliable product that helps users deploy models, monitor them, a/b test them, measure model drift, and so on. In a world with more and more tools for MLEs, we’re excited to bet on a bold vision that’s focused on the people who are at the heart of building models — the data scientists.

We were thrilled to partner with Harry and Tom when they were starting Periscope Data ten years ago, and we’re even more thrilled to partner with them on their next journey. If you want to use the next generation of model deployment tools, check out Modelbit, and if you want to work on the generation of these tools, they’re hiring!

Our Investment in Modelbit was originally published in Susa Ventures on Medium, where people are continuing the conversation by highlighting and responding to this story.


Comments are closed.