Clog

Inspiration

Clog is inspired one Kate's (a member of the Clog team's) lived experience searching for a job in the tech industry. She wasted a LOT of time on great-sounding opportunities that were not real.

What it does

The platform enables the job-seeking community to help each other my feeding a machine learning model that can identify fake listings.

How we built it

We took some of the excellent Chrome extension developer tutorials from Google and repurposed them to build the user interface. The ML model is developed and run using Google's Document AI Workbench. Supporting backend services use Google Firebase.

Challenges we ran into

The most accessible platforms for building document classifiers don't seem to like HTML documents as an input, so we had to convert to PDF.

Accomplishments that we're proud of

We have a working ML model.

It's a great name.

What we learned

Chrome extensions are very easy to make.

We were surprised to learn that building a simple document classifier still takes a fair amount of developer and/or cloud platform knowledge.

What's next for Clog

Our Chrome extension is really more of a proof of concept than a working demo. All of the underlying backend needs to be built and wired in.

Built With

• chrome
• documentai
• google-cloud
• javascript