Systems | Development | Analytics | API | Testing

NERSC, the mission high-performance computing center for the U.S. Department of Energy Office of Science, is using ClearML as part of the AI infrastructure stack for Perlmutter, the upcoming Doudna supercomputer, and the broader American Science Cloud. Here is a look at what they are exploring and why it matters for AI for science at scale.

How open standards give security teams more control, not less.

The biggest operational bottleneck in property management isn’t a lack of technology. It’s the manual coordination required between systems, teams, and processes. Leasing coordinators paste data from the PMS into email threads. Maintenance supervisors scan spreadsheets to find overdue work orders. Accounting teams wait for someone to confirm a deposit before posting. Owner reports get assembled the night before a call because nothing triggers them automatically.

I had the distinct pleasure of hosting a Snowflake Summit ‘26 session with Agustin “Augie” Del Rio, CEO and Founder of Gallus Insights, an analytics platform tailored specifically for mortgage lenders. As we sat down to discuss the future of analytics, one core truth echoed throughout the room: the most ambitious AI goals live or die by the quality of the underlying data.

When Temporal launched, a lot of people had the same reaction: "We have queues and retries. We don't need this." (Temporal's own blog addressed this directly.) That reaction made sense. Queues solve queue problems and they do it well. What Temporal gave you was something different: a named execution context that survives a server restart and picks up from its last checkpoint. Not a better queue. A different abstraction entirely. If you built with it, you couldn't imagine going back.

Generative AI for QA is the use of large language models to accelerate the creation and analysis of testing artifacts — drafting test cases, summarizing requirements, and generating synthetic test data. AI agents extend that capability into multi-step autonomous workflows that plan, delegate, and execute testing tasks across an entire delivery pipeline. For SDETs, the shift is not about learning to prompt more cleverly.

The question isn't whether to use AI in QA. It's knowing exactly where to keep a human in control. The core risk: Over 75% of multi-agent failures are silent semantic errors that pass automated checks but violate business logic — detectable only by human inspection (Cemri, Pan et al., NeurIPS 2025). The division of labor: AI owns repetitive generation and execution; humans own risk analysis, requirement interpretation, exploratory investigation, and final sign-off. The operational discipline.

Every enterprise contract I’ve signed in the last two years has the same clause. “Vendor will not use Customer Data to train machine learning models.” Sometimes it’s a paragraph. Sometimes it’s a whole section. The language varies but the intent is identical: don’t feed our production data into your AI. I get it. I sign the same clause as a vendor. But here’s what’s been bothering me: that clause is a promise, not an architecture.