Z Combinator

Inspiration

I want to lower the barrier for new people who are starting to deal with infrastructure to be able to learn about big and complex infrastructures easily. I wanted tinkerers, founders, developers, and large teams to have better insights into their product/application.

What it does

first and foremost it is a useful tool in itself where it is describing a big and complex infrastructure with something as simple as a graph with vertices being the component or instance that the infrastructure is composed of and their relationships (or frankly anything other kind of a relation we would like) being displayed by the edges.

second it has a natural language query engine on top of it to power the simple static graph into a more usable tool.

third in and itself i explore a lot of AWS infrastructure concepts through what i have built using my tool that are applied in the real world and it gives a lot of insights into how a big web infrastructure of an application works internally and in the cloud.

How I built it

Using Terraform files to generate AWS (small, medium, and large scale) infrastructures in a very "infrastructure as a service" style. I used neo4j as a database to hold all the infrastructure data and given the nature of neo4j as a graph database it reveals interesting insights. I am now building a way to turn natural language queries into actual graph queries that can be integrated into the application.

Challenges I ran into

Before Terraform, I was looking into spinning up a swarm of docker containers each representing a piece of a component of an application infrastructure. There were way too many complications dealing with so many docker instances. Writing all of the infrastructure in one single config file was by far the biggest challenge that I ran into and I feel terraform simplified a lot of process (thanks to mentor Morgan)

Accomplishments that I am proud of

As a proof of concept i feel like it really makes a really simple idea realize its full potential with the natural language queries.

What I learned

I learned a lot about how infrastructures behave and interact with each other while also learning how they can be deployed in a real world setting.

What's next for Z Combinator

Adding a Language Model kind of a technology to dynamically handle natural language queries and convert them to Graph Queries which will expand even further what the users can make do with their infrastructures. Maybe again use language agents who can go in to the infrastructure and make changes.

Built With

• amazon-web-services
• cloud
• ec2
• javascript
• neo4j
• sagemaker
• terraform