Kosmas Karadimitriou: An AI Expert's Journey from Innovator to Founder

Technology is not just a career for Kosmas Karadimitriou—it's a journey, one that's as much about human potential as it is about algorithms. In an industry often cast in the gloss of the West Coast, Kosmas provides a fitting counterpoint, a seasoned technologist whose story begins not amidst California's high-tech hubs, but in the cradle of Western civilization—Greece.

A multifaceted professional, Kosmas has already left a lasting mark across the spectrum—from AI and machine learning (AI/ML), to data science and engineering, to influential roles in the executive sphere. With a track record boasting 11 patents for innovative algorithms and solutions, and crucial contributions to five successful startups, including industry giants like KAYAK and ZoomInfo, Kosmas's credentials paint the picture of a career shaped by innovative thinking and a thirst for technological advancement.

When Shopify unexpectedly shuttered its logistics and robotics divisions in May of this year, Kosmas, along with 2,000 highly skilled colleagues, found themselves in the corporate wilderness. But while the layoff was a shock, Kosmas did something admirable—instead of reeling, he rallied, launching his new company Nova Science Ventures and signing engagements within days of Shopify’s announcement. He saw it not as an ending, but as a transition—an unexpected turn in the road to be met with optimism, not despair.

Kosmas went from a bachelor’s degree in Physics at Aristotle University of Thessaloniki to a PhD in Computer Science in the United States. He spent two years in mandatory military service in the Greek Air Force, equipping him with discipline and determination that would serve him well in the tech arena.

Kosmas’s belief in technology's power to create a better future is contagious: “I dedicated my life to understanding not only how the universe works, but also trying to predict how the future will look, and how to use technology to safely get us there—making sure that our destination would be a utopia and not a dystopia.”

Kosmas’s commitment, perseverance, and optimism remind us that in every perceived setback, there can be opportunity—if only we know to look.

We were lucky enough to speak to Kosmas about his journey, the future of AI, and how technology can create a better world for us all.

Kosmas, you’ve had an expansive career in AI/ML dating back to 1999 and ranging from edtech to healthtech to water pipe failures to e-commerce. How have you seen the field change and evolve over this time?

Over the past 25 years we’ve witnessed the remarkable evolution of AI/ML from a niche academic research discipline, to a core capability of modern high-tech businesses across a broad range of industries. This transformation was catalyzed by a confluence of factors, including the exponential increase in computational power and data availability, the rise and rapid advancements of deep learning, and the democratization of AI/ML through open-source libraries. Back in the late '90s, the term "Data Science" didn't even exist, and business applications for AI/ML were very rare and difficult to implement. Today, these technologies are the main drivers of innovation and transformation in diverse fields, from edtech to healthtech, traveltech, agtech, fintech, e-commerce, etc. It's been an extraordinary journey and an exciting time to be alive and witnessing this evolution, particularly for people who directly work with AI technologies.

What is your vision for the future of AI in e-commerce?

AI can help transform e-commerce in several ways:

1. Hyper-personalization: AI will continue to improve our ability to understand customer preferences at an individual level. Machine learning models will be able to predict with greater accuracy the products a customer might be interested in, their preferred price points, their desired customer journey, and even the best times to engage with them.

2. Seamless Conversational and Voice Commerce: With the proliferation of voice-activated devices, customers will be able to order products, track their orders, and receive product recommendations through natural, conversational interactions with AI-powered assistants.

3. AI-Driven Supply Chain and Logistics: AI will increasingly manage supply chains, inventory, and logistics, using predictive analytics to anticipate demand fluctuations and optimize stock levels. This will not only reduce costs but also improve customer satisfaction, ensuring products are available and delivered on time.

4. Sentiment Analysis: AI can analyze customer feedback, reviews, and social media posts to understand customer sentiment towards products, services, or the brand as a whole, providing insights for improving products and services, and for better targeting marketing efforts.

5. Augmented Reality (AR) Shopping: AI can enable more immersive shopping experiences by integrating AR technology. Customers can 'try on' clothes or see how furniture might look in their homes before making a purchase.

There are also many other applications where AI/ML can provide a lot of value in e-commerce, for example, better fraud detection capabilities, price optimization, competitive analysis, automated customer service through intelligent chatbots, and so on.

How have your experiences at research institutions like NASA and the M.D. Anderson Cancer Center influenced your approach to problem-solving in the private sector?

Work in research institutions puts a lot of emphasis on the application of the rigorous scientific method. This is invaluable for companies in the private sector that seek to foster data-driven cultures that could help avoid potentially costly mistakes or inefficiencies. Decisions based on data are usually far better and more scalable than subjective or arbitrary decisions.

A statistician would always win in the long term over any gambler, and this is particularly evident in the private sector, where some companies may enjoy short-term success by being lucky to be in the right place at the right time, however, long-term success comes to those who use science, data, and rigorous math and engineering to steadily advance.

You have 11 patents in AI/ML applications. Could you talk about one of these patents that you're particularly proud of and its real-world impact?

I am particularly proud of the first patents I authored in my career as part of my work in the founding team of Zoominfo. These patents were about the usage of Machine Learning (in particular, Bayesian Networks) to solve real-life business problems such as analyzing and classifying the contents of a website, in order to extract meaning and information. I am proud of these patents because we did that work at a time when no Machine Learning libraries existed, and even the practice of training and testing ML models was not well developed or understood.

Essentially, we had to start from some generic and abstract math, and use C++ to code from scratch, train, and deploy a fully working ML system. Nowadays this type of work is quite automated and fairly easy to do using standard libraries, but back in early 2000 that was a groundbreaking effort, pioneering the usage of ML for solving real-world problems.

I am fascinated by your work as CTO of VODA.ai. How were you and your team able to use AI/ML technologies to predict water pipe failures? What were the challenges and outcomes?

AI/ML technologies thrive on data: given access to sufficient and good data, very accurate predictions are possible. However, the challenge with water utilities is that their data are often incomplete or inaccurate: older records exist only on paper or don’t exist at all, whereas even newer records can contain inaccuracies since the crews that make repairs are not necessarily trained for or even motivated to do accurate record keeping.

As a result, a very large part of our effort was into cleaning and preparing the input data. Once the data are in a good state, the problem becomes easier: combining the utility data (e.g. pipes, breaks, etc) with other publicly available data (e.g. soil, traffic, roads, etc) and feeding them into our sophisticated purpose-built AI engine can yield some very accurate predictions about the condition of pipes, i.e. essentially performing a virtual pipe assessment, ranking pipes by their likelihood of failure. This helps utilities prioritize pipe monitoring and maintenance, thus reducing the number of unexpected pipe failures and leaks that a city experiences.

As a co-founder and board member of Circular Genomics, how are you applying your AI/ML expertise in the field of Precision Medicine?

AI/ML is being leveraged to fundamentally transform the way we understand and treat diseases, specifically focusing on brain disorders.

In Circular Genomics we start by collecting and analyzing circRNA data, which is a novel and promising approach to finding biomarkers related to brain disorders. By harnessing AI's power, we can accelerate the analysis and interpretation of circRNA data, and combined with other patient data such as demographics, previous diagnoses, treatment history, etc, we can predict how a patient may respond to a specific treatment. This personalized approach allows us to move away from a "one-size-fits-all" model of healthcare to one that takes into account individual variability. It's a challenging journey, but the potential impact on patient care and outcomes makes it incredibly rewarding.

You have approached Shopify’s shutdown of its robotics and logistics divisions in an extremely positive and constructive way. Could you talk a bit more about what this series of tech layoffs means for the industry, and how you see it changing the ecosystem?

While tech layoffs like the one at Shopify might initially seem negative, I view them as transformative events that can reshape the industry landscape in many ways. At the heart of these transformations is talent mobility, which in the tech industry is particularly high.

One of the key impacts is the redistribution of highly skilled workforce. The sudden availability of top-tier talent means that this expertise can spread across the ecosystem, for example to smaller companies who might not have access to such talent otherwise. It's akin to planting seeds of innovation across the industry.

Additionally, layoffs often spur entrepreneurship. Many of the employees affected by these changes are experts in their fields and can choose this moment to start their own ventures, whether consulting firms, tech startups, or other initiatives. This can lead to the creation of novel products, services, and solutions that wouldn't have otherwise existed.

Despite the immediate hardship for those directly affected (and I was one of them!) I believe that tech layoffs can, in the longer term, diversify and strengthen the industry ecosystem, accelerating innovation and fostering growth.

On that note, tell us about your new company, Nova Science Ventures - what is your vision for the business?

This is my attempt into trying something new, providing consulting services and even funding new projects and initiatives centered around the latest advances in AI/ML. Large Language Models are an incredible technology that suddenly has solved so many technical problems, from computer reasoning to language translation to intelligent code completion to content generation, and beyond. Nova Science Ventures can help bigger companies engage with AI, or can fund startups to bring new ideas and AI applications to life.

The rise of Artificial Intelligence is a watershed moment for the world and civilization, arguably as momentous as the invention of the Internet or even electricity. Our world will change very rapidly and in unprecedented ways, and Nova Science Ventures is my way to follow closely and make sense of these advancements. It’s a very exciting (and perhaps a little scary!) time to be alive, we do not know how the world will be 5 or 10 years from now, but one thing is certain, it will be very different from today. Hopefully for the better!

Is there any one guiding principle that you live by?

One guiding principle throughout my career has been continuous learning and adaptation. Particularly in today’s rapidly evolving world, staying stagnant is not an option. This principle is not just about keeping up with new technologies or techniques; it's about evolving my thinking, challenging my assumptions, and constantly seeking out new perspectives.

A second principle is a deep-seated belief in the transformative power of technology to solve complex problems and improve lives. This belief has guided my choices, leading me to contribute to sectors as varied as e-commerce, water management, traveltech, edtech, and healthtech. It pushes me to use my skills in AI and Data Science to create not just business value, but also to make a positive societal impact.

These guiding principles—continuous learning and my belief in the power of technology for good—have not only shaped my career, they have also helped me navigate changes and challenges along the way.

And finally, as a closing thought, every time I encounter adversity in life I always remember and get inspired by that age-old Latin proverb: “If there’s no wind, row”. Ultimately, we are responsible for choosing and navigating our own path in life, and even when the conditions are less than ideal, we can still push forward and make a positive impact.

‍

Images courtesy of: Kosmas Karadimitriou and Circular Genomics