Why we invested in Arize AI: Repeat Trinity entrepreneur and a massive ML market
Arize AI is the perfect example of what we look for in our seed investments — a proven entrepreneur fixing a problem they have faced in an area that aligns with Trinity’s investment focus. We are delighted to be funding Jason Lopatecki and Aparna Dhinakaran’s new startup Arize AI, which provides monitoring and explainability for machine learning ops teams.
AI Infrastructure is the Next Big Wave
One of our core theses at Trinity is investing in AI infrastructure, a market that is expected to be worth $50.6B by 2025. Every large enterprise we talk to is investing significantly in ML resources and talent.
When we break down the AI / ML space, we see a few key segments. We made our first bet in the data preparation layer with an investment in Figure Eight (formerly CrowdFlower), which was founded by Lukas Biewald and recently acquired by Appen for $300M. A few years later, we funded Lukas’s next startup Weights & Biases, which helps developers prepare, iterate and manage models, and then deploy them into production. Now, we are placing another bet in the AI observability space with Arize AI. Observability is an area we know well, having made an early investment into New Relic (NYSE: NEWR) for application and web infrastructure monitoring. We believe monitoring may be even more critical for AI and we constantly hear about how ML Ops teams are building their own solutions in house to tackle this issue.
Proven Second-time Trinity Entrepreneur
It’s an investor’s dream to back a talented entrepreneur across multiple startups. I had the privilege of seeing Jason’s expertise firsthand when he was a founding executive at TubeMogul. At the time, Aparna was also leading the fraud detection and mitigation efforts there. Trinity led the Series A for TubeMogul which ultimately went public in a highly successful IPO and was later acquired by Adobe.
Jason’s superpowers are understanding his customers, building and iterating on products to solve their pain points and recruiting great technical talent. Before a single line of code was written for Arize AI, Jason and Aparna had spoken to over 100 potential C-level executive customers as they were forming the business. This level of hustle and customer obsession is exactly what we look for in entrepreneurs.
Solving a Vexing AI Production Challenge
Arize AI was born from Jason and Aparna’s frustration from constantly debugging and explaining AI issues to their teams and customers at their previous companies.
When Jason was deploying AI models at TubeMogul in the advertising space and the model failed, it reflected poorly on the whole company. Jason needed a way to practice effective AI performance management, but there weren’t adequate technology solutions on the market that addressed the problem. So he did what any fearless entrepreneur would do: He built it himself.
Meanwhile, Aparna was working through similar frustrations on the Uber Marketplace Forecasting team where she was helping build out machine learning infrastructure to deploy thousands of models. She led the development of MLMS, Uber’s platform for model lifecycle management which is now an integral component of Michelangelo, Uber’s Machine Learning Platform. In an effort to bring a solution to enterprises for the pain points she observed, Aparna used her impressive technical chops to create her own startup called MonitorML, which was a part of the Y Combinator S19 batch.
Having crossed paths over five years earlier in their work together at TubeMogul, Jason and Aparna reconnected and realized they could build a category-defining company by joining forces and starting Arize AI.
With a $4M seed financing round and a simultaneous acquisition of MonitorML, Arize AI has the fuel and talent to take this market. We are excited that we are partnering with Foundation Capital GP Ashu Garg and former TubeMogul CEO Brett Wilson in this round, both of whom were prior EIRs at Trinity! I’m honored to be part of Jason and Aparna’s exciting journey ahead and offer the entire Arize AI team a warm welcome (and for some a welcome back) to our Trinity Family. Onward and upward!