Iterative.ai, an MLOps company developing data science and AI engineering workflows, today announced that it raised $20 million. The company says it’ll support the launch of its first commercial product, Data Version Control (DVC) DVC Studio, a dashboard aimed at simplifying machine learning model development based on data and metrics.

MLOps, a compound of “machine learning” and “information technology operations,” is a newer discipline involving collaboration between data scientists and IT professionals with the aim of productizing machine learning algorithms. The market for such solutions could grow from a nascent $350 million to $4 billion by 2025, according to Cognilytica. But certain nuances can make implementing MLOps a challenge. A survey by NewVantage Partners found that only 15% of leading enterprises have deployed AI capabilities into production at any scale.

Iterative, which was founded by ex-Microsoft data scientist Dmitry Petrov and entrepreneur Ivan Shcheklein, maintains a number of products designed to solve MLOps challenges including Continuous Machine Learning (CML), DVC, and the aforementioned Studio.

Above: A screenshot of Iterative.ai’s Studio product.

Image Credit: Iterative.ai

CML enables data scientists to manage machine learning experiments and track who trained models or modified data and when. They can codify data and models instead of pushing them to a Git repo and set CML to auto-generate reports with metrics and plots, building machine learning workflows using services like Amazon Web Services (AWS), Microsoft Azure, and Google Cloud Platform.

DVC is an open source version control system for machine learning projects that’s designed to make models shareable and reproducible by handling large files, datasets, models, and metrics as well as code. DVC connects these components via language-agnostic pipelines and leverages cloud storage, network-attached storage, or disks to store file contents. Full code and data provenance help track the metrics of every model, while push-pull commands move bundles of models, data, and code into production or remote machines.

As for Studio, which combines DVC and CML into a fully managed suite, it lets data scientists organize and navigate through multiple machine learning projects while creating teams, adding team members, and inviting them to experiment. Studio helps to visualize data and metrics through plots, trend charts, and tabular presentations and compare experiments. Studio also keeps code, data, and experiments connected, so that every change generates insights into how models can be improved.

Growing MLOps market

According to Iterative CEO Petrov, the advantage of MLOps is that it puts operations teams at the forefront of best practices within an organization. The bottleneck that results from machine learning algorithms eases with a smarter division of expertise and collaboration from operations and data teams — and MLOps tightens that loop.

“AI platforms and solutions are built outside of the software development technology stack. It creates a wall between ML researchers and software engineers. It prevents machine learning folks from using best practices and tools from software development. Our goal is to break this wall and create the best collaboration environment for both machine learning folks and software engineers,” Petrov told VentureBeat via email. “[As a result of the pandemic,] companies are paying more attention to automation. MLOps is becoming a more mature area and attracting more interest from companies.”

Iterative competes with Molecula, which is developing a cloud-based feature store for AI and machine learning workloads. Another leading rival is Domino Data Lab, a startup developing a platform focused on enterprises with large data science teams.

But Florian Leibert, a general partner at Iterative investor 468 Capital who also invested in the company, has confidence in Iterative’s go-to-market approach. Leibert is the founder of Mesosphere, an infrastructure startup based on the open source software Apache Mesos, which abstracts compute resources like CPUs away from machines.

Iterative claims that over 1,000 companies are using its tools and that its open source projects have a combined total of more than 200 contributors and 4,000 community members.

“Data, machine learning, and AI are becoming an essential part of the industry and IT infrastructure. Companies with great open source adoption and bottom-up market strategy, like Iterative, are going to define the standards for AI tools and processes around building machine learning models,” Leibert said in a press release.

468 Capital and Leibert led 15-employee San Francisco, California-based Iterative’s latest funding round, a series A, with participation from investors True Ventures and Afore Capital. It brings the company’s total funding to more than $25 million to date.