Galileo comes out of stealth to make it easier to develop AI models

- Advertisement -


As the use of AI becomes more widespread in the enterprise, the demand for products that make it easier more and more to check, detect and fix critical AI bugs. After all, AI is expensive – Gartner predicted in 2021, a third of technology vendors will invest $1 million or more in AI by 2023, and debugging an algorithm that has gone wrong threatens to inflate the development budget. Separate Gartner report found that only 53% of projects make it from prototype to production, presumably due in part to errors—a significant loss when the costs are calculated.

- Advertisement -

Tired of the high failure rate and the fact that secondary (if important) data preparation tasks, such as loading and cleaning data, Still occupy most of the time of data scientists – co-founders Vikram Chatterjee, Atindrio Sanyal and Yash Sheth Galileo, a service designed to act as a collaborative writing system for developing AI models. Galileo tracks AI development processes, using statistical algorithms to pinpoint potential system failure points.

- Advertisement -

“There were no specialized data processing tools for machine learning on the market, so [we] launched Galileo to build a stack of machine learning data science tools, starting with [specialization in] unstructured data,” Chatterjee told TechCrunch via email. “[The service] helps machine learning teams improve their datasets…by identifying critical datasets that may be underrepresented or erroneous, while at the same time being a comprehensive solution that encourages data scientists to proactively monitor data changes in production and address errors and gaps in their models from leaking into the real world.”

Chatterjee has a background in data science, having worked at Google for three years as a product manager on the Android team. Sanyal was a senior software engineer at Apple, focusing mainly on Siri-related products, before becoming a lead engineer on Uber’s AI team. As for Sheth, he also worked at Google as a full-time software engineer, leading the Google department. Speech recognizer Platform.

- Advertisement -

With Galileo coming out of stealth today with $5.1 million in seed funding, Chatterjee, Sanyal, and Sheth set out to build a product that could scale across the entire AI workflow—from pre-development to post-production—and for data modalities like text, speech and vision. Available in closed beta and built to be deployed locally, Galileo aims to streamline pipelines between teams using “autologgers” and algorithms that identify system issues.

Finding these issues is often a major challenge for data scientists. According to one recent survey (from the MLOps community), 84.3% of data scientists and machine learning engineers say the time it takes to detect and diagnose model issues is a problem for their teams, while more than one in four (26, 2%) admit that it takes them a week or more to find and fix problems.

“The discussion about machine learning within the enterprise has shifted from the question “What do I use it for?” to “How can I speed up, improve, and reduce the cost of machine learning workflows?” Chatterjee said. “Galileo …ensures the necessary rigor and proactive application of science-based methods at every step in the production of machine learning models… [It] leads to an order of magnitude improvement in how teams handle the messy, tedious task of improving their machine learning datasets.”

Galileo fits into the emerging practice of MLOps, which combines machine learning, DevOps, and data engineering to deploy and support AI models in production environments. The market for MLOps services could reach $4 billion by 2025. gradeand includes startups such as Databricks, DataRobot, Algorithmia, as well as incumbents such as Google Cloud and Amazon Web Services.

While investor interest in MLOps is on the rise, cash doesn’t necessarily lead to success. Even today’s best MLOps platforms cannot solve all of the common problems associated with AI workflows, especially when business leaders cannot quantify the ROI of these initiatives. The MLOps community survey found that convincing stakeholders that a new model is better, for example, remains a problem “at least sometimes” for over 80% of machine learning professionals.

Chatterjee points to Kaggle CEO Anthony Goldbloom’s investment in Galileo — Goldbloom led the seed round with The Factory — as a sign of the company’s differentiation. Chatterjee says Galileo currently has “dozens” of paying customers, from Fortune 500 companies to early-stage startups — revenue that Galileo plans to use to triple the size of its 14-person team by the end of the year.

“Galileo has focused on turning the painstaking task of machine learning data validation to simplify it and deliver data mining quickly,” Chatterjee said. “The user only needs to add a few lines of code.”

To date, Galileo has raised $5.5 million in venture capital.


Credit: techcrunch.com /

- Advertisement -

Stay on top - Get the daily news in your inbox

DMCA / Correction Notice

Recent Articles

Related Stories

Stay on top - Get the daily news in your inbox