From Raw Data to Beautiful UX: Dynamic Email Templating in Serverless FinTech

Day 57. Because users don’t want to read database dumps over their morning coffee.

When engaging with financial APIs such as Plaid, one swiftly recognizes that accounting principles diverge from conventional human reasoning. In typical API outputs, a positive numerical value represents an expense, whereas a negative numerical value denotes a credit or refund.

When my AI Financial Agent disseminated its daily communication, it informed users of their “Latest Activity: -4.22€” pertaining to an interest payment. This resulted in confusion for both the AI and the users.

On Day 57, I undertook a comprehensive revision of the notification infrastructure. I developed a dynamic HTML generator within my AWS Lambda environment. Prior to the dispatch of the email via Amazon SES, the Python worker intercepts the payload, segregates the arrays, applies the abs() function to negative floating-point numbers, and dynamically incorporates CSS classes according to the type of transaction.

Moreover, I enhanced the AWS EventBridge orchestrator. Rather than relying on multiple cron jobs, a singular hourly pulse activates the Lambda function. It verifies whether the user has surpassed their daily budget (which triggers an immediate text message via Amazon SNS). If the server’s UTC time coincides with 09:00 AM local time, it disseminates SQS messages to initiate the resource-intensive AI email generation.

Sophisticated backend engineering facilitates aesthetically pleasing frontend interactions.