On Wednesday, Seattle-based machine learning platform OctoML announced the closing of its $28 million Series B funding round. Addition led the Series B, which brings OctoML’s total funding to $47 million since its launch in 2019.
Machine learning is often touted as a revolutionary technology that will change the way computers “think,” but there’s a big gap in how many machine learning models are created and how many actually get used. According to OctoML, about 90 percent of machine learning models don’t make it into production.
“Machine learning has become mission-critical in virtually every industry, yet getting models to production remains labor intensive, slow and cost-prohibitive,” CEO and co-founder Luis Ceze said in a statement. “While ML spend is on the rise, 90 percent of models don’t make it to production. This is because improving model performance without sacrificing accuracy requires endless manual optimizations and fine tuning, especially given the growing stack of ML software and hardware backends.”
Software teams have to jump through a number of hoops in order to get their machine learning models to the point where they work when deployed. This involves making sure the algorithms compute fast enough, don’t break when scaled, work at the edge level and so on.
So OctoML has created a machine learning acceleration platform that automates the process of optimizing, benchmarking and deploying machine learning models. Essentially, this takes a machine learning concept and makes it a reality.
This tool is quite valuable to companies trying to deploy their machine learning technology. Several major tech companies — including AMD, Qualcomm, Bosch and Microsoft — are already using OctoML’s core product, Octomizer, since its launch six months ago. Early results show that Octomizer can bring a 30x boost in performance improvements, according to the company.
“By using ML to optimize ML, we reduce the optimization and tuning time by orders of magnitude. A 30x boost in performance translates to 30x savings in compute cost,” Ceze added.
Octomizer is still in its early access stage, but OctoML is allowing more companies to sign up for its waitlist. This new funding will help the company get closer to a larger product release.
In order to do so, OctoML will also grow its team aggressively. The company currently has about 45 employees, and it is now aiming to double its team. This involves making several hires across engineering, sales, customer success and marketing. A handful of these roles are already open.
OctoML is working on building momentum in order to become a recognizable brand in the machine learning space. But in the meantime, the company has received the full support of its investors.
“When we first met Luis [Ceze] and the OctoML team, we knew they were poised to transform the way ML teams deploy their machine learning models,” Lee Fixel, the founder of OctoML investor Addition, said in a statement. “They have the vision, the talent and the technology to drive ML transformation across every major enterprise.”
