🚲 Hi, my name is Teddy Kurita!

I’m a software engineering student at Carleton University and will be interning with IBM for the next year! More specifically, I will be working on the IBM TechXchange Community team, so I’m excited to start this journey of documenting my experience and connecting with and learning from fellow technologists (and perhaps cycling enthusiasts).

Think 2025 was happening during my first week, which presented the perfect opportunity to learn about IBM while orienting myself to the new environment. The opening keynote featured IBM’s CEO, Arvind Krishna, who highlighted IBM’s collaboration with Scuderia Ferrari HP. I was immediately captivated by the demonstration of how AI with watsonx can be used in a real-life context. As someone who is new to IBM products, I tend to compare IBM’s models and agents against external LLMs. However, this comparison overlooks the time and resource costs that come with using LLMs.

What piqued my interest was the approach that AI agents took compared to LLMs. LLMs (I find) take a “maximum capability in all areas” approach, whereas AI agents are designed to have a specific skill such that they are very efficient. From the keynote, we learned how Ferrari is using the watsonx platform to process more than a million data points per second generated by Formula 1 race cars. By utilizing the speed of AI agents, the mobile experience provided for Ferrari fans benefits greatly by being able to provide real-time information about the race. Learning about this use case of AI agents and watsonx Orchestrate has taught me a lot about a design philosophy for AI that I hadn’t really considered before.

IBM watsonx Orchestrate is a generative AI and automation solution that automates tasks to simplify complex processes. While familiarizing myself with watsonx Orchestrate, I realized that it’s directly related to software architecture styles that I’ve seen in my studies (watsonx Orchestrate uses the multitenant architecture style, which shares many similarities with the microservices architecture style). It's frequently difficult to notice and understand underlying architecture styles of software systems. The difference with watsonx Orchestrate is that it was presented in such a way that the architecture is apparent, which is a refreshing change. This knowledge will help challenge my views when implementing or using AI in future projects or work.

If you want to learn more about watsonx Orchestrate, watch Ritika Gunnar’s keynote on AI assistants.

Watching Think 2025 was the perfect way to start my internship: I learned a lot about IBM products and the impact of IBM on a global scale. I’m excited to learn more across the course of the year and participate within the community!

See you next time! 🐻