“AI on production. What’s the catch?”
During the talk, I will explore some of the challenges you face when deploying AI-based services worldwide. Of course, it’s nice to see high numbers in fancy LLM benchmarks, but when it comes to commercialization, other metrics emerge from the void. I will share some stories related to releases, model (over)complexity, testing, and maintenance. Everything will be based on our lab’s experience in commercializing Samsung AI services such as Galaxy AI on Mobile, Vision AI on TV, and Bixby Voice Assistant. I hope this will be a good opportunity to look at AI development from a slightly different perspective.
Bio
I’m Marcin — a mathematician by trade turned AI team lead. I run a crew of engineers, linguists, testers, project managers and researchers, bridging research, code and language to build AI-powered systems that actually get used. Over the past few years I’ve led and shipped projects in voice assistants, data analytics, machine translation, opinion mining, information retrieval, and other AI features powering Samsung Galaxy and related products. I love rolling up my sleeves, experimenting, and mentoring folks across disciplines — whether in data science, software, or linguistics. Outside of work, I’m a literature lover (I love Joyce, Beckett, Saramago), deeply curious about philosophy, and passionate about languages and Portuguese culture — all of which shape how I think about technology and people.
