Venue: Day n Nite, 2-4 Observatory Court, Green View Court, Tsim Sha Tsui, Hong Kong
Date: 26-Sep-2024 Thur
Time: 19:00 - 21:00
Rundown :
19:00 Registration & networking
19:30 - 20:00 Building Real-time Next Generation AI Applications with Kafka and Flink - Jeffrey Lam, staff solutions engineer, Confluent
20:10 - 20:20 Introduction of Hacktober Fest 2024 - Open Source Hong Kong
20:30 - 21:00 InstructLab: A new way for everyone to customize LLMs collaboratively using an open-sourced way - Peter Ho, senior solution architect, Red Hat
21:30 Closing
Remark:
Please reserve the event and contact OSHK Exco when you arrived.
For this event, we will give away a limited number of dining coupons while stocks last.
The minimum consumption in the restaurant is one drink, so please comply.
A small amount of snacks will be provided at the start of the event.
Thanks again to Red Hat / Confluent for the F&B sponsorship.
Open Source Hong Kong has the final decision on event arrangements. For this event, we will give away a limited number of dining coupons while stocks last.
The minimum consumption in the restaurant is one drink, so please comply.
A small amount of snacks will be provided at the start of the event.
Please keep the place clean, thank you for your cooperation.
Thanks again to Red Hat / Confluent for the food sponsorship.
Open Source Hong Kong has the final decision on event arrangements.
注意事項
請在此活動留位及到當日到現場和開源香港同事登記
是次活動，我們有限量的餐飲券乙張送出🍻，送完即止。
餐廳的最低消費為一杯飲品，敬請配合。
開場時，現場會在提供小量的小食。
再次感謝 Red Hat / Confluent 餐飲贊助。
請保持場地清潔，多謝合作
開源香港對活動安排有最終決定權。
—
Session Details
Building Real-time Next Generation AI Applications with Kafka and Flink - Jeffrey Lam
Description: With the hype of Gen AI worldwide, how to continuously enrich trustworthy data streams to quickly build, deploy, secure and scale real-time AI applications has become a core competitive advantage. Kafka and Flink form a natural foundation due to their fit for processing real-time and batch data at scale and their huge adoption in both the commercial and open-source community.
InstructLab: A new way for everyone to customize LLMs collaboratively using an open-sourced way - Peter Ho
Description: Large Language Models (LLMs) initially started as proprietary systems, keeping their model weights and training data closed. Recently, open-source LLMs have begun to gain traction, offering more transparency. However, even with open model weights, the training process remains tightly controlled by the providers, preventing community members from contributing changes directly. The challenge is to create a model where the open-source community can participate in refining LLMs, allowing contributions like custom tuning and knowledge updates. InstructLab, developed by Red Hat and IBM, aims to address this issue. It is an open-source project that provides a command-line interface (CLI) and methodologies to facilitate community-driven fine-tuning. This session will introduce InstructLab and demonstrate how it empowers community members to contribute to model improvements.
