An Introduction to Modern AI: Neural Networks, Generative Models, and Practical Limits

April 8, 2026

• Title: An Introduction to Modern AI: Neural Networks, Generative Models, and Practical Limits
• Speaker: Artem Kholodov | Software Developer, IMAX
• Date: Wednesday, April 8, 2026
• Time: 4:00 PM
• Room: LH3058

Summary

Since the introduction of ChatGPT to the general public in November 2022, AI has exploded into public view. However, while the capabilities of AI models grow year after year, a lack of understanding of the technology, including belief in many falsehoods, still pervades some areas of public discourse. An incomplete grasp of the fundamentals, coupled with unbridled optimism, has left many with an unrealistic view of the capabilities of generative AI.

In this talk, I will introduce the fundamentals of machine learning and neural networks, exploring the history and mathematics of how these models evolved from limited binary classifiers to being able to recognize images, play games, and generate text. I will then discuss these models in the context of productivity tools and "Agentic AI" by providing a nuanced examination of the capabilities and fundamental limitations of this class of models. By drawing on my professional experience and building upon core machine learning concepts, I will provide a realistic and clear picture of what these tools can and cannot achieve in practice, recent advancements in the field, and where we might expect these capabilities to go in the future.

Biography

Artem Kholodov is a professional software developer with over a decade of experience in the tech industry. A graduate of the University of Waterloo, he holds a Bachelor of Computer Science and a Master of Data Science and Artificial Intelligence. His professional experience includes data engineering, large-scale distributed computing, and the design and implementation of corporate agentic workflows. Having worked professionally before and during the AI boom, he witnessed the rapid transformation of the software engineering landscape firsthand, and by combining this professional experience with a rigorous academic background is well situated to bridge the gap between AI theory and practical considerations of its use as a productivity tool.