MLOps platform Domino Information Lab nabs $100M

Domino Data Lab, a San Francisco, California-based supplier of MLOps options, at present introduced that it raised $100 million in a sequence F funding spherical led by Nice Hill Companions and an expanded partnership with Nvidia, which additionally participated alongside Coatue Administration, Highland Capital Companions, and Sequoia Capital. The proceeds, which carry Domino’s whole raised so far to $228 million, will likely be put towards product growth and platform growth as the corporate seems to be to bolster its international buyer base, in keeping with CEO Nick Elprin.

MLOps, a compound of “machine studying” and “data know-how operations,” is a more moderen self-discipline involving collaboration between information scientists and IT professionals with the goal of productizing machine studying algorithms. The marketplace for such options may develop from a nascent $350 million to $4 billion by 2025, according to Cognilytica. However sure nuances could make implementing MLOps a problem. A survey by NewVantage Companions discovered that solely 15% of main enterprises have deployed AI capabilities into manufacturing at any scale.

Domino, which was based in 2013 by Chris Yang, Matthew Granade, and Elprin, presents a collection of machine studying mannequin administration instruments that assist fight points like idea drift, or when the statistical properties a mannequin is making an attempt to foretell shift over time. The corporate’s digital workbench permits engineers to leverage current instruments in monitoring, reproducing, and evaluating experiments whereas discovering, discussing, and reusing work in a single place.

“As enterprises advance their adoption of machine studying, they more and more require an enterprise-grade, open platform to orchestrate and handle these workloads — and Domino completely meets this rising market want,” Nice Hill Companions’ Derek Schoettle mentioned in a press release.

MLOps platform

Utilizing Domino, clients — which embrace greater than 20% of the Fortune 100 — can spin up interactive workspaces on the {hardware} of their alternative and scale to extra highly effective compute sources if crucial. The platform offers a built-in bundle supervisor to prepare the libraries and instruments tapped all through a challenge, and versioned datasets to trace information used throughout mannequin coaching and testing. Domino’s reporting options let admins schedule experiences to be generated and delivered to stakeholders routinely, whereas information pipelines deal with duties to maintain fashions updated.

On the mannequin operations aspect of the equation, Domino lets clients deploy fashions as on-demand APIs or export fashions for deployment on different infrastructure. It could possibly detect information drift and monitor the efficiency of fashions within the wild whereas alerting engineers to people who underperform, with a registry that reveals the standing of fashions at a look. Domino additionally retains tabs on computing and measures the enterprise affect of mannequin APIs and apps because it surfaces challenge well being when it comes to each progress and potential roadblocks.

Domino says that Nvidia’s funding will allow it to develop product performance to broaden the accelerated computing capabilities of its platform. This consists of validating the Domino platform for Nvidia AI Enterprise, a managed assortment of software program instruments designed to speed up machine studying workloads in order that Domino can run on Nvidia-certified techniques from OEM {hardware} suppliers. Since 2020, Domino has provided licensed enterprise MLOps software program for Nvidia DGX techniques as a member of Nvidia’s DGX-ready software program program.

“AI and information science are new workloads that demand a full-stack answer — one with instruments that simplify growth and deployment for purchasers,” Nvidia head of enterprise computing Manuvir Das mentioned in a press release. “Our partnership with Domino Information Lab displays our dedication to enterprise MLOps. With Nvidia AI Enterprise integration into the Domino platform, clients will be capable to simply combine superior AI instruments into their conventional information middle infrastructure.”

Domino, whose clients embrace Dell, Allstate, UBS, Bristol Meyers, ConocoPhillips, and Lockheed Martin, beforehand raised $43 million in a funding spherical co-led by Highland Capital Companions and Dell Capital. The corporate competes with a number of startups within the multibillion-dollar MLOps market, together with Iterative.ai, Comet, Weights and Biases, and DataRobot, which in July closed a $300 million financing tranche at a $6.3 billion post-money valuation.

To maintain up with the competitors, Domino recently added new capabilities that combine with Git repositories in a means that makes it doable to extra simply monitor all facets of mannequin experimentation. The corporate additionally added a function that makes it doable to run a number of sandboxed environments concurrently to enhance productiveness, in addition to a brand new set of applied sciences — Domino Mannequin Monitor — designed to forestall fashions from degrading.