The secret to getting AI on the fast track A blue man representing an AI against a blue digital background.

What’s your CIO hiring for these days? Knowing the answer — and how to be the leader who has those skills — might just be the ticket to the next phase of your career. More than half of the CIOs surveyed by IT analyst firm Gartner in 2021 said that they plan to boost staffing to manage machine learning and AI initiatives.

The urgency is because AI workloads have traditionally required specialized IT infrastructure. That is, until now. AI and data science workloads can now run on accelerated mainstream servers in enterprise data centers, which makes it simple for IT professionals to support these new applications.

About the author

Anne Hecht, senior director of Enterprise Solutions at NVIDIA.

Running AI in the enterprise is now easier than ever, yet to understand project requirements IT still needs to know about the language spoken by AI experts. Luckily, the fastest way to learn the AI lingo is also the fastest way to find success in AI projects: include your company’s data scientists when you’re developing plans for your enterprise AI infrastructure.

The enterprise AI VIP

When it comes to AI, an enterprise’s most important users are its data scientists. They’re serious AI experts, and they know that they need powerful servers with accelerated computing capabilities to get their work done on deadline. Data scientists crunch enormous data sets, so a lot of compute power is needed to iterate, refine and revise their AI models before they can be put into production.

Having data scientists at the table when you develop your AI infrastructure plans will help you define their use cases so you can prepare for your organization’s long-term AI ambitions. You’ll also get live feedback to help you eliminate options that don’t fit their requirements.

Circle the wagons

Inviting data science teams to join planning for AI might seem like an obvious step on the road to success, but since the domain is new for many organizations, it’s not always a common practice.

In fact, an IT manager at a leading healthcare organization recently mentioned how he was asked to invite three people he didn’t know to a meeting. Someone wise must have been planning the guest list, because it turned out that those three colleagues were on the data science team.

With their input, the IT team was able to jump past dead-end ideas that wouldn’t align with what the data science team needed to develop their AI project, and then get it running in production. Having everyone aligned meant that IT could focus on solving the compute and infrastructure problems, while the data science team advised on their application and workflow requirements.

Avoiding the shadows

When IT and data science teams aren’t connected, “shadow AI” becomes a serious risk. This happens when AI developers find ways to get the infrastructure they need outside the organization’s IT infrastructure.

Many data scientists would happily focus on their own domains and leave the infrastructure to IT, but their work also requires powerful systems. By nature, data scientists tend to be creative problem solvers. They might take matters into their own hands when needed.

To find a way to get their work done efficiently, data scientists might turn to cloud computing outside IT’s usual control, or run their workloads on unsecured personal systems. These workaround solutions add both cost and risk for enterprises, and can be easily avoided by making sure IT is working closely with their AI experts.

Easy enterprise AI on traditional infrastructure

Until very recently, AI infrastructure was almost always siloed away from the mainstream servers IT uses to run most business applications. This made AI and its infrastructure the domain of specialized experts, but it’s now possible for AI to run on industry-standard servers managed by enterprise IT teams.

To easily deploy AI on traditional data center servers, look for purpose-built enterprise AI software that makes it easy to deploy these advanced workloads on hybrid clouds with scalable high performance. The right tools can provide the AI frameworks and tools needed by data scientists and AI developers, with integration into your existing infrastructure ecosystem.

Become an AI IT expert

AI is a tectonic shift in technology that’s similar to the change organizations had to navigate when first getting onto the internet. This new era of AI is creating demand for new skills, new software and new systems. With those come new opportunities for IT professionals.

AI workloads may be novel for many enterprises, but this presents a perfect opportunity for IT teams to develop new expertise that not only increases their value to their organizations, but also helps grow the careers of individual IT professionals.

By working with your data science experts, IT can find success the first time and avoid potentially costly failures with purchases that don’t hit the mark. In fact, looping in your data science experts might just turn out to be the foundation of the next chapter of your IT career.

social experiment by Livio Acerbo #greengroundit #techradar