Jane Edwards 
- Tim McNamee, Cloudera’s public sector account manager, has highlighted how private AI could help agencies manage costs and data governance
- Cloudera’s private AI approach includes AI studios, accelerators and synthetic data tools
- McNamee said the goal is to let departments build their own AI models
Tim McNamee, enterprise account manager for the public sector at Cloudera, said private AI could help government agencies unlock the benefits of artificial intelligence while avoiding common cost and data barriers.
What Challenges Does Private AI Address for Agencies?
In a video interview posted on GovLoop on July 27, McNamee explained that cost unpredictability is one of two major hurdles slowing AI adoption in government.
He said the other challenge centers on data gravity and data sovereignty, noting that agencies hold petabytes of sensitive data, including tax records, health data and intelligence information that often cannot be moved to the cloud. He added that agencies should not respond to these barriers by slowing down innovation, but by moving innovation closer to their data.
Why Is Data Quality Important for AI Initiatives?
McNamee highlighted data governance and data quality as central considerations for agencies pursuing AI initiatives.
“One of the more important things, and if you look across government and the policies they’re putting out for their AI strategies, is really around data governance and data lineage,” he said. “So really understanding the truth of your data and making sure you have quality data to run AI against to prevent the risk of inaccurate outcomes.”
What Does Cloudera’s Private AI Approach Include?
Cloudera’s private AI approach is built on four components: an AI platform with frameworks, models and services; data management software, including an open data lake house; accelerated compute, including GPUs and networking; and secure enterprise-grade servers and storage.
During the interview, McNamee highlighted Accelerators for Machine Learning Projects, or AMPs, and Cloudera’s AI Studios. McNamee said the AMPs are designed to solve a range of enterprise problems beyond fraud, waste and abuse, including image recognition and voice-to-text. On the AI Studios, McNamee pointed specifically to Cloudera’s Synthetic Data Studio, noting that it removes personally identifiable information from training data to help reduce bias and exposure risk.
McNamee said Cloudera’s goal is to let people beyond traditional AI teams build AI models that support their own work.
“What our goal is to do is to not only let people use the traditional AI tools, but allow, say, pharmacies or other departments to create AI models that can help them with their job,” he noted.
How Does Cloudera AI Help Speed Up Agency Adoption?
To advance private AI, Cloudera has partnered with leading server providers to offer Cloudera AI, a turnkey, single SKU solution or appliance. McNamee said this approach helps agencies avoid a common delay in government IT: buying software and then waiting months for servers to arrive.
“So by providing this AI solution, agencies can really more rapidly get time-to-value with the solution,” McNamee said, adding that he believes the approach will change the AI industry more broadly.
