Machine learning has become an indispensable technology for running a business, but the process of building a model in it still requires iteration and experimentation. Comet, on the other hand, created an overall platform for finishing models from ideas to products, and announced the $ 50 million Series B on November 18th. Prior to this, Series A was $ 13 million in April.
Previous investors Scale Venture Partners, Trilogy Equity Partners, and Two Sigma Ventures participated in the OpenView-led B-round. According to Crunchbase data, the company’s cumulative funding is close to $ 70 million.
According to co-founder and CEO Gideon Mendels, the product can be used on any platform, whether it’s a laptop, the cloud, or an on-premises cluster. “Comet manages and optimizes the entire machine learning lifecycle, from experimental follow-up to model production monitoring, so it’s a platform that empowers data scientists and accelerates development for machine learning engineers.” Mendels says.
According to Mendelsso, the approach has been successful, with the company’s annual recurring revenue quintupling in 2021 and 150 companies including Uber, Zappos and Etsy using Comet. Also, according to OpenView partner and lead investor Mackey Craven, he was attracted to Comet because it creates effective products for emerging markets with great opportunities. .. “What we are seeing now is a rare combination of outstanding founding teams that can be the core of large, lasting market opportunities that can fully support them, and the changes in that market. The reason for that change is that new entrants are creating and capturing the value we create in the big markets of the future through the creation of new markets or technology shifts, “says Craven.
Currently, the company has 50 employees and comes from nine countries on four continents. The plan is to have 100 people in 2022. Diversity and inclusion are important parts of the company’s value system, according to Mendelssohn. “That’s the core of our corporate culture, and 35% of our employees are still minorities, and we’ll take that into account in future employment,” he said.
The company’s new product, Artifacts, is a data versioning tool that works and works like document versioning. It is used by data scientists to keep track of data changes.
According to Mendelsso, “When you’re working in a machine learning pipeline, Comet’s Artifacts allows you to automatically version each snapshot of your data. Every time you make a change, you get that version. “. There are many advantages, but the main one is that the data scientist can see how the data used for training the model has changed, and the data of the model at the time of training can be compared with the data of the final production.
Image Credit: yucelyilmaz / Getty Images
[To the original text]
(Sentence: Ron Miller, Translation: Hiroshi Iwatani)