Not an ORM but a query builder. So developers are encouraged or sometimes forced to think about SQL and database layer, which is a good thing.

We use Node for API server, Lambda, and some utility tools. The biggest reason is developers can share the same language with front-end. And Node.js is mature/active enough to be a decent startup choice.

Mature OLTP database choices. Less fragmented than MySQL

Manage infrastructure as codes. Native AWS solution so it has better support to AWS resources than Terraform, also can leverage AWS Business Support.

Authentication as a service. Instead of rolling our own user authentication flow for yet another time, it's better to leave the security concern to a service specialized in this field.

Native integration with Atlassian products (Jira and Confluence). Bitbucket Pipeline is clean and easy.

Docker serves as the default way of deployment packaging. And with many other tools dockerized, it's essential for the team to be familiar with Docker and use it when the time is right.

Serverless is the future. And AWS Lambda is the most mature FaaS out there. AWS SAM makes it easy to package Lambda as micro-apps.

Scale well from simple startup use cases to complicated enterprise situation.

GIF! (mic drop)

New and very popular. Less legacy to deal with compared to React. Great documentation. Easy to get started.

Powerful all-in-one monitoring solution as a service. Good integration with AWS. Very affordable price for small-scale startups.

Neat data transformation tool and only focusing on data transformation. Fitting the popular ELT paradigm. Small but active community. Easy to get help on their Slack from maintainers.

Data warehouse solution that fully separates compute and storage. Better management facility than directly using S3.