AI and ML: Simplifying Radical Innovation for the Future of Business
The speed of AI: Redefining efficiency
Today, with supercomputing and AI, teams can accomplish tasks in seconds that were impossible a few years ago. This powerful combination has allowed organizations to reimagine the boundaries of possibility—not just in medicine and finance but also in sports, advertising, customer support, and other data-driven sectors. That’s according to Dr. Rangan Sukumar, a distinguished chief technologist for HPE GreenLake, whose job includes helping HPE “build the world’s fastest supercomputers.”
Creating an architecture for the impossible
HPE supercomputers—some with the capacity to complete one billion billion (10^18) calculations in a second—are designed to transform machine learning (ML) and artificial intelligence (AI) into practical, user-friendly tools for organizations of any size, Sukumar says. Since these technologies are so much more efficient than manual methods of collecting, analyzing, and understanding data, they can expand human potential and productivity, Sukumar states.
How AI changes the way organizations approach innovation
According to Sukumar, AI can manage mundane, cumbersome, and possibly hazardous tasks within an organization and help teams do more with less effort.
In addition, AI tools can help organizations speed complex workflows past process bottlenecks, bridge gaps in multi-team communications, and deliver actionable insights across teams without sacrificing precision or slowing with scale.
Sukumar states that AI helps teams accomplish this by analyzing data with high speed and precision while allowing people to focus on what they do best—the pattern and trend discovery required for creative problem-solving. For example, when insights are readily available across an organization, decision-makers can build more effective data-driven strategies. When scientists can reduce research time without compromising quality, they may develop and test prototypes faster, bringing new solutions to the marketplace more quickly.
Collaborating to build future tech
“I meet with CTOs (chief technical officers) and CEOs, and we figure out what they’re trying to do for the organization,” says Sukumar. “We take that interaction with the leadership and convert it into innovative ideas, collaborative projects, and technology demonstrations that become HPE products they’ll need perhaps three years from now.”
Key to HPE GreenLake’s collaboration with global teams is the company’s attention to how AI can best fit into organizations immediately and as they continue to modernize over time. Tools that are challenging for teams to use are often underutilized, so HPE GreenLake focuses on user experience in tandem with end goals like efficiency and scalability. Sukumar says people need robust and frictionless tools to integrate AI into how they naturally communicate, strategize, and build new solutions for their unique business challenges.
“Data is at the heart of AI’s capacity to transform organizations, and the ‘intelligence’ part of AI depends on large language models (LLMs) having the right inputs to learn how to ‘think.”
DR. RANGAN SUKUMAR
CHIEF TECHNOLOGIST FOR HPE GREENLAKE
How organizations are training systems to think like people
According to Sukumar, organizations are leveraging the power of AI and machine learning that ‘thinks’ like an intelligent team. Systems powered by AI and machine learning can be programmed to intuitively route workflows to focus on tasks that deliver the most benefit while conserving and optimizing resources through, for example, the management of calendars or light and temperature controls. AI tools can also detect anomalies that are humanly impossible to identify, Sukumar says, while simultaneously fighting cyber attacks and prioritizing the delivery of insights connected to urgent needs.
Data is at the heart of AI’s capacity to transform organizations, says Sukumar, and the ‘intelligence’ part of AI depends on large language models (LLMs) having the right inputs to learn how to ‘think.’ The training of LLMs to execute code functions—which render decision recommendations and automate actions in IT and other systems—can be resource-intensive and highly complex to manage. According to Sukumar, 80% of enterprise data is unstructured, making data sets even more challenging to compile, analyze, and transform into insights with AI. Structuring that data and establishing relationships between data points often demands additional computing power and human resources for success.
“There’s no AI without the right data.”
According to Sukumar, because every organization’s needs are unique, insight-rich data makes AI more impactful with every input and interaction with the teams using it. That data informs how Sukumar supports companies as they match their AI strategies to their long-term goals. By helping teams discover how, where, and when to use AI to get value from their data, HPE GreenLake plays a pivotal role in transforming theoretical ideas about accelerating innovation into practical, market-ready applications.
Yet, as the adage goes, AI ‘is’ what it ingests: data. For organizations like financial service institutions (FSIs), sports teams, or medical research facilities, the rules for deciding what constitutes clean data can determine whether an AI model will fuel innovation or distribute a false narrative. HPE GreenLake conducts rigorous data integrity assessments to ensure that AI projects rely on timely, relevant, and comprehensive truths as they are built and launched.
Unlocking the right tools at the right time
HPE GreenLake helps organizations fast-track large-scale AI project development by providing them with the high-capacity data storage and data curation solutions necessary for effectively training LLMs, even when they have massive amounts of unstructured data to manage.
In addition, Sukumar says HPE GreenLake offers clients the building blocks required to innovate and develop their models more efficiently by giving them access to various options – like software-as-a-service (SaaS) and functions-as-a-service (FaaS)—on demand. This allows organizations to solve the intensive workflow and workload problems that AI model development often incurs by letting them use the features and services they need on demand from HPE GreenLake’s ecosystem of AI development tools.
Sukumar says he challenges his teams to ask, “‘What makes an AI model ‘stick’ within an organization, go into production, and stay there for a significant time and provide value?’”
DR. RANGAN SUKUMAR
CHIEF TECHNOLOGIST FOR HPE GREENLAKE
Connecting AI, innovation, scale, and efficiency
Good AI models can deliver powerful insights that give organizations an edge. However, those coveted bespoke models can be challenging to scale since they can break when faced with problems or use cases they’re unfamiliar with. The effort to get those models right can be daunting and can derail innovation strategies even after an organization has experienced some success in launching an AI project.
“When you think about scale, most people think it just means having the world’s best AI model on the leaderboard with the most parameters and intelligence packed into it,” Sukumar states. “We think of scale differently, especially for customers aiming to grow sustainably.” Sukumar says he challenges his teams to ask, “‘What makes an AI model ‘stick’ within an organization, go into production, and stay there for a significant time and provide value?’”
According to Sukumar, AI models survive only if they “reach a critical mass where the costs of training, fine-tuning, and developing inference show up as ROI.” “To sustain AI practices, you need to generate multiple ROI use cases that then fund themselves,” Sukumar says. “Typically, you pick the use case with the most potential, which then pays for the next step.”
From the lab to the playing field: AI is transforming how we think about strategy
Like research labs, sports teams are highly complex organizations. A team may succeed or fail based on how well it can extract, analyze, and react to new data, whether it’s a shift in league standings or the likelihood of a player being out for a season after an injury. Cloud-based AI tools make disparate data actionable in real time, so coaches and medical staff can access relevant insights instantly on any cloud-connected device.
“In terms of use cases, we’ve been able to predict injuries based on gait, sudden drops in speed, and other metrics while the athlete is on the field,” Sukumar says. “One interesting use case involved a baseball player who came back from injury but could not pitch as well as he used to,” Sukumar states. “By analyzing 3D data using AI from video feeds, we discovered that his shoulder’s angle had changed.”
As in other use cases, Sukumar says the speed of AI data analysis is critical. “Injuries happen on the field, and if you don’t act within the next 10 minutes, you could extend downtime by several days,” Sukumar states. “Quick decisions based on different types of images, such as X-rays or MRIs, can determine if immediate surgery is needed or if rest is sufficient—this can reduce player downtime.”
For Sukumar, the way clients have successfully used AI to launch radically innovative strategies comes down to collaboration—and an ecosystem of AI tools available for the task.
“To succeed in AI, you need end-to-end support,” Sukumar states. “This includes training, fine-tuning, inference, and the colocation of data and AI.” According to Sukumar, HPE supports all of this with experts, curated tools, and an open ecosystem.
paid and
presented by
Learn more about HPE GreenLake
HPE GreenLake enables organizations to transcend traditional operational limits and pioneer new territories in problem-solving and decision-making. Discover how HPE GreenLake can unlock AI's capacity to transform business outcomes.
Discover MoreSHARE THIS STORY