N Things I Want to Do with Snowflake in the Financial Industry

My company name is a secret. I am a member of a Fintech company in the financial industry.
My name is also off-limits. In the Snowflake community, I'm known as Nun-san.

I'd like to speak on the theme: "Things I want to do once we introduce Snowflake."

1. I want to eliminate data silos
   My company has many silos.
   First, there are many databases. There are various different database services.

The reasons are:

• The company provides multiple services
• The timing of service launches varies
• As a result, the database service that seemed optimal at each specific time was selected

These include on-premises databases, databases running on cloud resources, cloud-managed databases, etc.

There are two data lakes, depending on how you count them. The enterprise-wide data lake and the BI-focused data lake are separate. These are scheduled to be consolidated into one this year.

Since there are many silos and management is difficult, I want to unify them.

2. I want to separate compute resources and storage.
   Being in the financial industry, we close the books at the end of the month and calculate the flow of money from the previous month starting at the beginning of the next month.
   We calculate debits and credits using SQL, but I've heard it takes several days.
   I believe there are several reasons why it takes time, but one is likely that compute resources are fixed, which leads to long calculation times.

   Therefore, I want to use Snowflake, where you can change the warehouse size—setting it to XL at the beginning of the month and around L size for other periods.

3. I want to perform data masking
   It has been two to three years since I changed jobs and joined my current company. One thing that surprised me after changing jobs was
   that "personal information is raw data."

   I will not touch upon the work details. I will get in trouble with the company.
   I will speculate on the background of the design:

   • Masking takes up resources. They do not want to allocate resources.

4. I want to utilize AI
   Since the current model is machine learning-based, I want to use LLM.

5. I want to utilize Streamlit
   We are short on engineers. The same goes for application engineers.
   I want to create Streamlit applications.

That's all. Thank you for your attention.

Reference image for silos