Baseten Gives Data Science and Machine Learning Teams the Superpowers They Need to Build Production-Grade Machine Learning-Powered Apps

Print Friendly, PDF & Email

Baseten formally launched with its product that makes going from machine learning model to production-grade applications fast and easy by giving data science and machine learning teams the ability to incorporate machine learning into business processes without backend, frontend or MLOps knowledge. The product has been in private beta since last summer with well-known brands that have used it for everything from abuse detection to fraud prevention. It is in public beta at this time.

“It’s clear that the performance and capabilities of machine learning models are no longer the limiting factor to widespread machine learning adoption — instead, practitioners are struggling to integrate their models with real world business processes because of the enormous engineering effort required to do so. With Baseten, we’re reducing this burden and accelerating time to value by productizing the various skills needed to bring models to the real world,” said Tuhin Srivastava, co-founder and CEO of Baseten.

Over the last decade, there’s been enormous progress in advancing the capabilities of machine learning, driven primarily by new model architectures and the ever-decreasing cost of compute. But the critical step of integrating models with real-world business processes is still a lengthy, expensive process that prevents the majority of businesses from seeing a return on machine learning  investments. While a typical machine learning model may take just a few weeks to train, building the infrastructure, APIs and UI so that the model can be used by businesses can take more than six months and requires additional resources in the form of MLOps, backend and frontend engineers.

This is a problem that Baseten’s co-founders Tuhin Srivastava (CEO), Amir Haghighat (CTO) and Philip Howes (Chief Scientist) encountered first hand at Gumroad. There Haghighat was the head of engineering and Srivastava and Howes were both data scientists who had to learn to become full-stack engineers so they could use machine learning to detect fraud and moderate content. The systems they built at Gumroad are still in use and have screened hundreds of millions of dollars of transactions to date.

The trio founded Baseten so that data scientists don’t have to learn to become full-stack engineers in order to build web applications for their machine learning models. Baseten lowers the barrier to usable machine learning by enabling data science and machine learning teams to incorporate their machine learning  models into production-grade applications within hours instead of months. With Baseten, data science and machine learning teams can easily serve their models, build backends and frontends and ship applications that solve critical business problems including operations optimization, content moderation, fraud detection and lead scoring. 

Customers on Baseten:

  • “Baseten provides an easy way for us to host our models, iterate on them and experiment without worrying about any of the DevOps involved,” said Faaez Ul Haq, head of data science at Pipe.
  • “Baseten gets the process of tool-building out of the way, so we can focus on our key skills: modeling, measurement and problem solving,” said Nikhil Harithas, senior machine learning engineer at Patreon.
  • “Baseten allows us to take any need that we have and build a custom SaaS solution for that very specific use case — it’s amazing,” said Ryan Delk, co-founder and CEO at Primer.
  • “Baseten provides us with all of the speed and control of self-serving our model deployment, without any of the annoying config, infra and health checks,” said Daniel Whitenack, data scientist at SIL.

Analysts on Baseten:

  • “One of the top reasons that 90% of machine learning models never provide value is that businesses struggle to turn their finely tuned machine learning models into practical apps,” said Hyoun Park, Chief Analyst at Amalgam Insights. “Our research shows that the average successful machine learning project has an ROI of over 500%. By accelerating the time to translate models into apps, Baseten has the opportunity to unlock the value of machine learning for the enterprise and to allow companies that have to fully realize the competitive advantages created by machine learning.”
  • “The iterative lifecycle of building and operating machine learning models requires a multifaceted team,” said Kevin Petrie, Vice President of Research at Eckerson Group. “Baseten aims to help data scientists reduce their reliance on stakeholders such as DevOps engineers. This creates the opportunity to streamline and accelerate the deployment of machine learning models.”
  • “With highly operationalized platforms and pre-built or automated outcomes increasingly available, the act of building and deploying a machine learning model is rapidly becoming a mainstream endeavor across enterprises of all sizes. And yet, according to Omdia research, even though nearly two thirds of US companies are investigating or building pilot use cases, only nine percent have been able to bring those efforts through to production and even fewer (six percent) have done so at scale across the business,” said Bradley Shimmin, Chief Analyst AI platforms, Analytics and Data management. “What’s holding them back? Often, a simple impedance mismatch between developing a working model and integrating that model within the context of business apps at the point of action. With the ability to rapidly create APIs and embed machine learning models directly into shareable apps, Baseten promises to minimize this mismatch and accelerate time to value.”

Baseten Raises $20 Million in Seed and Series A Funding

Baseten also announced that it has raised $8 million in seed funding co-led by Greylock and South Park Commons Fund and $12 million in Series A funding led by Greylock. Baseten is using the funding to expand its engineering and go-to-market teams.

Greylock General Partner and Baseten Board Member Sarah Guo said: “Despite the broad understanding that AI has the capability to revolutionize business, most organizations struggle to drive real ROI from their machine learning efforts, stymied by the high upfront investment required. Baseten radically reduces the time, specialized expertise, cost and cross-team coordination required to successfully ship machine learning apps to production. Its end-to-end platform frees data science and machine learning teams from grunt work and empowers them to spend more time innovating and iterating to maximize impact. The Baseten team has experienced this pain first-hand, and that authenticity and care shows in the solution they’ve designed. We’re thrilled to partner with them to democratize access to the revolution in machine learning.”

Other participants in the seed round include AI Fund, Caffeinated Capital and angel investors Lachy Groom (ex-Stripe), Greg Brockman (co-founder and CTO of OpenAI), Dylan Field (co-founder and CEO of Figma), Mustafa Suleyman (co-founder of DeepMind) and DJ Patil (ex-Chief Data Scientist of the United States Office of Science and Technology Policy).

Other participants in the A round include South Park Commons and angel investors Lachy Groom, Cristina Cordova (ex-Stripe), Dev Ittycheria (CEO of MongoDB), Jay Simon (ex-President of Atlassian) and Jean-Denis Greze (CTO of Plaid).

Sign up for the free insideBIGDATA newsletter.

Join us on Twitter: @InsideBigData1 –

Speak Your Mind