Software Spotlight: Baseten
Company Snapshot
Founded: 2019
Employees: 20
Funding: $20M
Valuation: NA
Stage: Series A
Locations: San Francisco
Company Overview
Baseten makes it easy to incorporate machine learning into applications.
Tell Me More
Baseten enables software engineers to create production-grade ML applications without the backend, frontend, or MLOps knowledge.
Historically, operationalizing ML has required tremendous coordination. Beyond the sophisticated data science and product work, data teams need to implement pre and post-processing data workflows, work with DevOps and backend engineers to deploy and service the model, integrate the model into existing systems and processes, and equip the end users with sufficient context to interpret and accept or reject the ML predictions. This is no easy task and could take upwards of 6+ months to go live.
Having jumped through hoops in previous engineering roles, the team at Baseten is attempting to reduce the burden, and accelerate time to value by productizing the various skills and measures needed to bring models to the real world.
The company is solving this problem with its own ML Application Builder. It enables teams to quickly set up the backends, frontends, and MLOps to utilize ML models faster. Baseten allows organizations to leverage public and proprietary machine learning models by integrating them with custom business logic, and designing powerful web apps for customers, all without requiring any infrastructure.
Market Opportunity
Over the past decade, there’s been enormous progress in advancing the capabilities of ML, driven primarily by new model architectures and the decreased cost of compute. However, integrating models with real-world business processes is still a lengthy and expensive endeavor.
This, along with the fact that ML projects are already expensive severely limits the total ROI businesses can achieve by implementing ML solutions. While ML models may only take a few weeks to train, building the accompanying infrastructure, APIs, and UI that allow the models to be used by businesses can be time consuming and expensive.
While businesses can see the benefits in deploying ML to their applications, productizing ML has become increasingly difficult. In January 2022, LinkedIn released its annual data on worldwide trends in hiring. Last year, the role of “artificial intelligence practitioners” came in dead last (no. 15) on the company’s list of jobs on the rise. But on this year’s list, which expanded to 25 roles, it ranks machine learning engineer as the 4th fastest-growing job in the US. However, over the same time frame, job descriptions for ML engineers are seemingly becoming more lax, with the average years of experience dropping from 6+ years to 4+ years. Still, supply for ML engineers is constraint as many employers look at educational prerequisites as check-box for entering ML.
This confluence of factors all bodes well for Baseten whose goal is to broaden access to production-level ML regardless of domain expertise.
Why I like the company
Baseten was born out of the idea that machine learning will have a massive impact on the world, but it’s still difficult to extract value from sophisticated machine learning models. By creating an automated interface and backend infrastructure to deploy ML models, Baseten has the opportunity to broaden access to ML capabilities for all companies.
With Basten, users can build production-grade, workflow-integrated ML apps for use cases like integrating fraud models into mortgage approvals, creating a classification model from the Baseten “model zoo” to power a content moderation queue, or leveraging pre-built models to extract data from unstructured content. Previously, these solutions would have required multiple engineering teams. There have even been companies whose entire goal was to automate these exact processes.As mentioned in Baseten’s market opportunity, the need for production-grade ML is growing quicker than the number of ML engineers. The lack of ML engineering talent underscores the need for end-to-end modularized ML products to help users quickly deploy models, without all the un-sexy headaches that come with it. By allowing technical users to embed ML models into their applications, Basten is further democratizing access to enterprise ML.
Lastly, the team has extensive experience with the problem set as the founders dealt with this exact headache at Gumroad. Additionally, the rest of the team hails from promising tech companies where they held leadership roles at companies like Clover Health, Gumroad, Uber, Twitter, Radar, and more.
Basten’s Hiring Corner
Remember, Baseten is still early-stage, so if you don’t see a role that fits your skillet, reach out to learn more!
Similar Companies
Reminder
As always, you can find the abbreviated list of companies we’ve already talked about here. Below are links to the previous posts