Arize AI is applying machine learning to some of technology’s toughest problems. To continue with its mission, the company announced $19 million in Series A funding.
Battery Ventures led the round with participation from existing investors Foundation Capital, Trinity Ventures, The House Fund and Swift Ventures. The new round comes over a year after the company came out of stealth with $4 million in seed funding led by Foundation Capital.
Arize was co-founded by CEO Jason Lopatecki, a former TubeMogul executive, and chief product officer Aparna Dhinakaran, who previously built machine learning infrastructure at Uber and joined the company when Arize acquired her company Monitor ML, where she was CEO.
The company touts itself as “the first ML observability platform to help make machine learning models work in production.” Its technology monitors, explains and troubleshoots model and data issues.
“At the beginning of the year, we talked to hundreds of businesses deploying machine learning and having the same problems,” Lopatecki told TechCrunch. “Most of their investment was going into building better models and getting them out, but no one had any software to help with the issues.”
Companies use data to build models that they use to automate decisions, but without visibility to see if the models are working or not, it’s difficult to determine if the models are accountable, fair and responsible when implemented in the real world, Dhinakaran added. Arize can be integrated into a company’s AI systems within 30 days to detect performance and bias issues and show customers how to fix it.
Between its seed and Series A rounds, the company secured enterprise customers like Adobe and Twilio. It also scaled its team and went from building a product to having it “massively deployed” in fintech, healthcare, insurtech, adtech and retail institutions.
The opportunity for the Series A came as the company was inundated, Lopatecki said.
“Being a second-time founder, you can feel when the product is taking off, and for us, we could spot where we have that influxsion,” he added. “We want to double down on the product side and get the solution out to more people and get it into more hands.”
As such, the funding will go into product development, expanding industries and use cases and growing its team of 40. He expects to double that quickly, especially as the company continues to see 100% annual recurring revenue growth each quarter and a doubling of its customer base.
Dhinakaran forecasts that in five years, Arize’s technology will be deployed into every AI system.
“Every single top machine learning team is going to have that visibility on if their model is doing well and if it is unbiased,” she added. “It’s not just a red or green light of modeling, but enabling practitioners to serve up issues and fix them.”
As part of the investment, Dharmesh Thakker, general partner at Battery Ventures, is joining the Arize AI board. His firm mainly invests in business-to-business software, and Thakker, in particular, oversees infrastructure investments.
Every three months, the firm picks a new theme. In this case, his team had heard from portfolio companies that there wasn’t the tooling available for deploying models and monitoring them. They looked at about nine companies, including Arize, and the more they got to know the company, they decided it had the best vision and leadership.
He sees the future of machine learning observability being a combination of a well-designed product and instant gratification. Customers don’t want to wait, and even though another company might have all the features, because Arize AI focuses on observability and can quickly show value and vision is what makes them stand out.
“Being an engineer myself, I look for founders that have felt the same pain, and in this case, Jason and Aparna have felt the pain because observability was missing,” Thakker added. “We also look for leaders who can hire great people. Not only do they feel the pain, but they rallied this A-plus team around them.”