HeatWave Japan User Group. Q&A with Masataka Narita and Tatsuya Naito
Q1. We hear about GenAI every day. How can we leverage HeatWave GenAI specifically? What makes HeatWave GenAI unique?
The big difference is that the entire GenAI process can be completed within customer’s HeatWave. There are concerns to when you use GenAI by providing your company specific data that you do not want to disclose to external services.
In a case when a manufacturing customer wants to predict with GenAI and AutoML by log data for maintenance, AutoML first detects anomaly statistics from the log to determine which process is problematic. This task can be completed with just several SQL functions.
After that, if you ask in natural language like “What is the main problem with this log? Summarize in two sentences,” HeatWave GenAI can generate a clear, relevant answer without manual analysis by data engineers.
Since all of this work is done within the HeatWave service, there is no need to provide your highly confidential data to a third-party GenAI service.
HeatWave can make predictions using real-time transaction data without ETL. Tapping into real time data in highly secured environment – that makes HeatWave quite different from other GenAI services.
Q2. Heard that you have launched the world’s first HeatWave User Community as Oracle/MySQL partner. What kind of community activities do you do?
As a MySQL partner company since 2004, we Smart Style have been providing consulting and support for databases, mainly MySQL. Since May 2018 we have engaged with reselling, implementation, support, and operation of Oracle Cloud Infrastructure (OCI) as an Oracle Japan Cloud Sales Partner
Yes, we launched the world’s first user group in Japan in March 2023 called HeatWavejp. HeatWavejp is a user community where participants can learn about the benefits of HeatWave and share know-hows.
Aiming to reinforce connections between participants we provide a communication hub using Slack, hold online/offline meetups and study sessions every other month.
Q3. Who can join HeatWavejp?
We welcome participation not only of HeatWave users but also those who are just interested. Through communications of use cases and latest tips, members can exchange higher quality information casually.
Q4. What would most attract newly joined members in your community?
It would be our bi-monthly meetups, where community members do LTs and share results from HeatWave testing. In December we hold “Case Study Festival” meetup where we present multiple new case studies. You can check past event reports here.
You can view the famous Mexican startup company Aiwifi telling their experience to HeatWavejp members.
Q5. Is HeatWave Lakehouse associated with HeatWave GenAI?
Yes, quite. There is a strong relationship. Each HeatWave features released to date is the foundation for GenAI as well, and Lakehouse works as an important component to import unstructured data.
HeatWave Vector Store processes your company’s internal data, such as the log data mentioned above, within HeatWave services so that the log can be used as knowledge base in the RAG.
Users will be able to get more accurate, context-specific answers without moving confidential and valuable data to a separate external vector database.
Q6. Can AWS users leverage HeatWave GenAI as well?
Currently HeatWave GenAI, AutoML, and Lakehouse services are available on AWS in some regions like Tokyo. These days, many systems are already built on AWS. We believe that GenAI that runs within AWS end-to-end, will be a great choice for customers.
Q7. GenAI is the biggest update for HeatWave this year. Is GenAI actively discussed at your meetups?
Two months after HeatWave GenAI was launched, we customized the RAG of HeatWave GenAI to extend LLM by putting our own data, and presented the results of our verification at our meetup.
Several days before the meetup, a typhoon was approaching Japan. We used Japanese weather news on it, translated into English and let GenAI generate a summary in Japanese.
On top of that, we deliberately gave incorrect information to the RAG. The result was, as you know, it gave us a wrong answer.
At the end of the session we shared that HeatWave GenAI does not support multibyte characters, and that possibilities of being misinformed in the case of using wrong information – Users need to check the authenticity on their own.
Q8: Any recommendation to other groups on how to setup working and successful MUG?
We experiment every day to make things better, but I think loving the product and being passionate about technology (Yes, it is about HeatWave!) makes a difference.
Having fun is the most important in continuing the community activities.
Resources:
mySQL Heatwave Japan User Group (in Japanese)
HeatWavejp (MySQL HeatWave Japan User Group) You Tube Channel
…………………………..
Masataka Narita, MySQL Consultant, Smart Style.
Narita is a MySQL consultant at Smart Style, one of Japan’s leading MySQL partners. After experiencing support and operations for Oracle Database and MySQL, he is currently engaging with design, implementation, and consulting services with MySQL.
Narita posts technical information about MySQL and HeatWave through blogs.
Tatsuya Naito, HeatWave Japan Community Manager, Smart Style
Naito joined Smart Style in 2015 and has engaged in consulting sales of database solutions, mainly with MySQL. In March 2023, he launched HeatWavejp, the world’s first user group for HeatWave, in Japan and keeps leading the community and supporting operations of the community without resting.
Sponsored by MySQL/Oracle.