How to train your own LLM ๐งโ๐ป & networking ๐ค
Talk and live demo in Shibuya on fine-tuning and compressing your own lightweight LLM to run offline, followed by networking.
- When
- Where
- Samโs Bar Tokyo, 25-7 Maruyamacho (4F) Shibuya Tokyo , 4th FloorยทIn person
- Region
- Kanto (Tokyo)
- Organizer
- Tech Community Japan
- Language
- EN
- Source
- Meetup
Summary
Software engineer Ovidijus Parsiunas walks through building a personalised, lightweight AI model step by step. Starting from an off-the-shelf model, the talk covers how fine-tuning teaches it new information and how compression lets it run on any computer without an internet connection, ending with a live demo of a model answering questions about custom data.
The presentation, demo and Q&A take about 30 to 40 minutes, and the rest of the evening is networking at Sam's Bar Tokyo in Shibuya. It is aimed at developers and anyone curious about AI. The fee is 900 yen paid in advance or 1,000 yen on site, plus one drink order at the bar.
About the community
Tech Community Japan runs regular meetups in Tokyo for software engineers and people curious about technology, mixing short talks by practitioners with networking, alongside small-group coffee chats. Events are casual and mostly held in English around Shibuya.
#llm#fine-tuning#model-compression#on-device-ai#live-demo#networking