It was 2023 and Debanjan Saha had just given a presentation about artificial intelligence (AI) at one of the world’s leading financial institutions when he got into an Uber. The driver keenly asked about the technology. It was at that moment that he realized: AI had finally crossed the rubicon from science fiction to science fact.
With popular applications like ChatGPT capturing the public imagination, there’s been a marked shift toward the democratization of AI. So much so that Saha labels the times we’re in as the “iPhone moment for the AI industry.”
Saha has spent his entire illustrious career preparing for this “iPhone moment,” from his early forays in startups to his work at tech giants like Amazon Web Services (AWS) and Google. It’s in his DNA to think big and take risks.
Now, as CEO of AI company DataRobot, Saha helps companies with what he calls the “CICD” — continuous integration and continuous delivery — pipeline of AI. DataRobot has been delivering easier access to AI since its inception in 2012. The company itself has evolved along with the technology. Originally started with a focus on using automation to create AI models, it has since added MLOps and generative AI.
“When you think about DataRobot today, it really is a platform to solve business problems with AI,” Saha told The Forecast during the 2024 .NEXT conference in Barcelona, Spain, where he participated as an inaugural member of the Nutanix AI Partner program.
Putting Customers First
After graduating with a bachelor’s degree in computer science from the prestigious Indian Institute of Technology in Kharagpur, India, Saha pursued graduate studies in the same field at the University of Maryland, where he completed his master’s and doctoral degrees. The university subsequently named him to its alumni Hall of Fame. Along with his many patent applications and awards, his memberships at prestigious societies like the Institute of Electrical and Electronics Engineers (IEEE) and the Association for Computing Machinery (ACM) are a testimony to his impressive achievements.
Saha’s career arc from startups to large companies has given him a ringside seat to the breathtaking pace at which technology has evolved over the past few decades. He has used that experience and his insider knowledge of the data game to shape DataRobot into a cutting-edge AI company that has an empathetic, customer-first approach.
“Bringing AI to business is not easy because you need to speak AI and the business language, and it’s very difficult to find people who speak both,” Saha said.
But DataRobot does. One of the first companies to join the Nutanix AI Partner program, it manages the AI pipeline from nose to tail: building data models; testing, validation and compliance needed for governance; and deploying and managing solutions from a single pane of glass.
The Current State of AI Infrastructure
Today, Saha sees an impressive investment in AI infrastructure, which he calls “the plumbing for a new era.”
But he strikes a note of caution, too. Actual boots-on-the-ground AI implementations are still not at a volume that reflects the technology’s promise, he said, attributing the disconnect primarily to two challenges.
The first challenge is the value gap — a difference between the investment going into infrastructure and the actual output. Tangible AI applications and use cases that deliver ROI for business are still trailing behind.
“People are spending a lot of money training models, but ultimately those models have to solve some business problems, have to make things more productive, or predict things more accurately. Business value has to justify the investment,” he said.
“You need people who understand the business and the business challenges that they are trying to solve, as well as people who understand AI and how those two things can be put together.”
Because trust in AI-generated results and models is still shaky, the second challenge is the confidence gap. Saha said people are creating many prototypes, especially with Gen AI, but they need to be thoughtful and mitigate risk.
“It's not difficult to put together a chatbot that answers some questions,” he said.” It's hard to guarantee that they are going to answer the question correctly. There are things we need to do in terms of managing the risk and creating a framework where you can define and associate risk with various different types of use cases and handle them with the right level of mitigation, compliance, testing, and validation.”
DataRobot is helping customers bridge these gaps by drilling down on AI’s fundamental building blocks. They provide a platform that automates and simplifies the machine learning process so customers can build and launch AI applications with confidence.
Saha has been in the data trenches long enough to have seen AI’s expansion from predictive models, mostly under the purview of data scientists, to its democratization for everyday use through generative AI models.
You need both, he said.
“It’s like a human being who needs both the left and the right brain: You need generative AI and predictive AI together to solve many use cases,” he explained, citing as an example a retail client of DataRobot that uses predictive AI to forecast inventory levels and generative AI to create personalized customer coupons.
The Future of AI
Saha believes generative AI will evolve as new media sources are added.
“We have reached the limits in terms of written text,” he said. “We have a lot of video, audio and images that will drive the next round of model training.”
Now it’s time to shift the focus to using it to create business value.
“Because AI has become a company-wide priority, CIOs are getting more involved,” he said.
Because compute-hungry AI requires massive amounts of energy, sustainability is one of the challenges that CIOs are most struggling with, especially as data center charges affect the balance sheets, Saha noted. “There’s a lot of work going on in making chips more power-efficient. There are also a lot of things you can do at the software and application layers to address the challenge,” he said. “For example, not every question needs to be answered by a large-language model.”
Companies will have to optimize both the software and hardware stack to reduce AI’s energy footprint, suggested Saha, who cited value creation and security as other chief concerns. Any enterprise that wishes to succeed with AI in the era of cloud computing must therefore move from “vision to value” by focusing on rapid experimentation and identifying projects with low-hanging fruit that can easily deliver ROI on AI investments, he said.
“On average, 54% of AI projects make it from pilot to production, meaning that nearly half of AI projects fail. Massive investments in AI struggle to deliver tangible value as they are too often met with brittle, hand-stitched tooling, data and organization silos, and governance and compliance blind spots — making scalable, continuous and repeatable success nearly impossible,” Saha wrote in a March 2023 blog post. “I believe that focusing relentlessly on value is what will enable us to bring the full potential of AI to life.”
Editor’s note: Learn more about the Nutanix AI platform, GPT-in-a-Box, for jumpstarting AI transformation with optimal infrastructure that delivers control, privacy, and security. Learn more about the Nutanix AI Partner program.
Poornima Apte is a trained engineer turned technology writer. Her specialties include engineering, AI, IoT, automation, robotics, climate tech and cybersecurity. Poornima's original reporting on Indian Americans moving to India in the wake of the country's economic boom won her an award from the South Asian Journalists’ Association. Poornima is a proud member of the Cloud (the sky, not the tech kind) Appreciation Society. Find her at wordcumulus.com.
© 2024 Nutanix, Inc. All rights reserved. For additional information and important legal disclaimers, please go here.