How to train your own LLM 🧑‍💻 & networking 🤝

About this event

Most LLMs live in huge data centers. However, recent advances now allow models to be personalized and deployed on any computer with no internet connection. How do you teach a model new knowledge and how do you make it smaller? Join Ovi as he walks through the whole journey of building your own personalised lightweight AI model, step by step. This session is for everyone — developers and anyone curious about AI. 🎤 Presentation • Overview: We’ll start with an off-the-shelf AI model. From there, we’ll explore how fine-tuning teaches them new information and how compression allows them to run anywhere. We’ll finish with a live demo of a model that can answer questions about our custom information. Whether you're a developer or simply curious about AI, you’ll leave with a clear understanding of how AI models learn and common techniques for compressing them. • Format: Presentation + Live Demo + Q&A • Duration: About 30 - 40 minutes (the rest of the time will be a networking part) • Note: No equipment needed: speaker will handle all setup 🧑‍💻 About Ovidijus Parsiunas Ovidijus is a Software Engineer from the UK who recently moved to Japan. A Machine Learning enthusiast since 2017, he has implemented multiple AI-driven solutions and is an experienced presenter on AI/ML topics. He owns several popular Open Source libraries including Deep Chat, MyVision, and Agentic Airport, and is one of the few people worldwide to have contributed to Open Source every single day for the past 7+ years. He was also a former member of the Machine Learning Tokyo group, which he joined back in 2018. Here is his LinkedIn profile. https://www.linkedin.com/in/ovidijus-parsiunas-1a8a85106/ 📌 Fee • Pay in advance （Stripe）：900 yen (Pay here (https://buy.stripe.com/3cIbJ18fvg1V6HL5R11gs0g)) • (The payment link is available 1 day before the event.) • Pay onsite：1000 yen (cash or PayPay) (\*Plus one drink order is needed at the bar.) ⚠️ Notes • If you don't come to this event without cancelation, others might be prioritized to join next time. • If the event is canceled by the organizer, a full refund of Stripe will be provded. No refunds will be issued for participant cancellations.