In today’s data-driven landscape, a shift towards a cloud-centric approach in data and analytics charts a new trajectory for innovation and efficiency. Embracing the cloud for data storage, processing, and analysis revolutionizes accessibility and scalability. This approach not only streamlines operations but also fosters collaboration, empowers real-time decision-making, and fuels the development of advanced analytics models.
To give us more insights on this we have with us Venkatesh(Venki) Shivanna, Senior Technical Director, EADP DATA & AI, @ Electronic Arts, Inc, who is a seasoned leader with over two decades of expertise in digital transformation, specializing in deploying technology at scale to create significant value. His leadership focuses on centralizing data collection, leveraging data-driven strategies to solve complex problems, and enhancing product experiences for personalized gameplay through secure channels. A visionary in storytelling with data, he excels in crafting self-service reports and narratives, translating raw numbers into actionable insights. Venkatesh’s extensive experience includes designing and implementing enterprise-level data platforms supporting Business Intelligence, Artificial Intelligence, and Data Analytics, managing vast data volumes from various organizational systems. He actively champions data governance best practices and mentors engineering teams, fostering scalable, fault-tolerant, and sustainable self-service data layers.
The discussion centers on the evolving data science landscape and the renewed interest in Data Foundational concepts amidst Generation AI’s rise. It highlights the importance of a strong data foundation in the cloud-centric approach and examines challenges in shifting from traditional models to cloud-based AI applications. Cost considerations in contrast to flexibility are explored, alongside factors influencing decisions for data storage and application deployment in the cloud. Finally, it predicts the cloud’s role in propelling future data science innovations and Generation AI’s emergence.
AIM: How has your journey been in the evolving data science industry? What prompted you to engage in this discussion, particularly focusing on the resurgence of interest in foundational concepts amid the widespread desire for Gen AI?
Venkatesh Shivanna: We all know that Gen AI is the next big innovation in today’s world, and everyone is talking about that, but I always take a step back and see how a stronger data foundation can be built around all these things were Data and analytics are the core part of them. Again, the Gen AI boom is just the beginning i.e., tip of the iceberg if you think about it. It sounds more exciting. Nobody knows what Gen AI is going to turn into. It has unknown vulnerabilities. There are a lot of things that have yet to be proven. Industries are still in early exploration mode. But to see that becoming a reality may take a few years. The reason is simple and depends on how effectively foundational data platform will feed datapoints to train these next Gen AI models to predict results back to the customer in a more promised way without biased opinions. Again, we are indeed in a more exciting phase of our technological career. Things are moving fast with more powerful GPUs in the industry can now crunch multi–Peta bytes of data at scale.
IMO, it is either a small or a large enterprise. Building the foundational data platform is critical, and that is where the Cloud centric approach will come into the picture to realize how rapidly we can deploy the apps to support business needs.
Again, it all comes down to how effectively we run our data strategy along with the company’s business strategy to evaluate how the next gen technological advancements can help companies to be more competitive in the Marketplace to make right decisions to provide next best personal experience to the customers.
AIM: Does building a robust foundation, as emphasized, indeed define the transformative power within the data and analytics realm, particularly within the cloud-centric approach? How do you envision this shaping the future of the industry?
Venkatesh Shivanna: The cloud-centric approach is all about how quickly we can build the foundational layers at scale in rapid time. A 4-7 years ago, it used to take many months or sometimes years to build these platform engineering teams to collect the data, process it, feed it back to our customer’s needs. With Cloud Centric Approach that journey has been completely transformed into a much faster now. As said earlier, how fast we can collect and process the data in near real-time to understand the customer behavior is critical to reach them in no-time before they churn out of your product offerings.
AIM: What challenges prompted the widespread adoption of a cloud-centric approach for most AI applications today, considering the shift from premise, hybrid, or non-cloud models previously used?
Venkatesh Shivanna: It is all coming down to maintenance overhead, like how much time being spent managing your infrastructure with cost in-mind. Sometimes it becomes more overwhelming for a Data engineering or IT teams in the company.
Having a cloud centric approach, engineering teams can focus more towards fulfilling the business strategy needs on how they can build better data products in agile manner. I can say those days are gone when we used to build infrastructure for a year or two to address organizations’ multi-year business strategy directly impacting being less competitive in the marketplace.
The cloud centric approach indeed gives us all the luxury of using S*/P*/I*AAS at scale to support our business needs, gives us more power tools to build and support customer needs.
AIM: Is the flexibility offered by this approach offset by the cost factor? Has it discouraged many companies from fully embracing the cloud and constructing their applications?
Venkatesh Shivanna: Every business has its limitations. Costing is the main factor in today’s Macroeconomics world. Hence, it is critical for enterprises to pick hybrid cloud strategy to see how their Digital Data Assets can retrofits @ enterprise level with an ability to migrate to the new providers in no-time which can give them multi-cloud strategy to negotiate better pricing model for adoption.
In summary, at an Enterprise level. Multi-cloud strategy and a hybrid Cloud strategy are key to the success of digital transformation across the company’s operational business units.
AIM: How do you navigate the decision-making process for storing data and enabling applications in the cloud, considering a set of three to five key parameters? What factors weigh into your final decision-making regarding the choice of cloud services?
Venkatesh Shivanna: Most critical criteria is how stable the cloud solution providers are, since your business will be running on their cloud service and stability is key to you own success. If they go down, your business is out.
Second, see how the cloud providers are partnering with different vendors like software and hardware vendors. Are vendor agnostic or not is key for your company’s Data strategy to avoid any friction?
Third, data security, i.e., protecting your digital estate is key which is nothing but an abstract reference to a collection of tangible owned assets. In a digital estate, those assets include virtual machines (VMs), servers, applications, data, and so on running on the cloud. Just imagine some of your products running on tens of thousands of nodes and how cloud providers can effectively manage your estate in a secure manner.
Again, from security POV, if anything happens within your public/private cloud solution, it’s critical to assess how the cloud providers can handle data/system breach incidents to completely Obsolete compromised nodes as needed without impacting your day-to-day activities across regions and isolate as it happens in real-time to a run your business at scale.
Fourth, Data Compliance which are critical in today’s world to protect customer’s privacy data likes of PII, PCI or PHI to support regulatory needs for GDPR, SAR, COPPA & CCPA etc.
AIM: Is data more secure on the cloud or within private networks, especially considering the vast amounts of AI and data analytics applications?
Venkatesh Shivanna: It depends on how your cloud centric strategy are aligned with the business needs to adopt right approach to build systems @scale to collect data in a secure way through either public vs private API end points depending on client-side applications or backend microservices used within your enterprise eco system.
Hence, we have different options to use in the Markets like REST API to connect Client apps to microservice(C2S) using JSON or S2S where gRPC can comes in handy using Protobuff to achieve high-performance, binary format and tightly coupled data binding protocol used to send data in transmission to secure them.
Sometime client facing applications can come under DDOS attacks and it may result in flooding your consumption end point with more data. So, it is extremely critical how we build systems to manage edge cases and apply certain mechanism to bypass them using rate limits or IP based allow list and encrypt the good data in transmission at the same time and persist them with right access control in place at rest etc.
AIM: What is the next significant breakthrough you predict, leveraging the cloud’s capabilities in enabling data science applications, driving multiple innovations, and giving rise to Generation AI as a byproduct?
Venkatesh Shivanna: IMO, it is all about how organizations can leverage the goldmine of external data available in the open world marketplace likes of IoT sensors data and Computer vision data to use it for their competitive advantage along with their internal data about their customers mobility within their own Eco system & find ways to engage them more with more personalized experience to improve long-term revenue generation opportunities. I call them “Connected Data eco system running the world’s day-to-day activities.”
Finally, building context oriented meta systems around all these data being collected is extremely critical to the success of the GenAI needs to train the AI models to make this world a better place to live and secure it for the future generation to take it forward.