hike | Devpost

How hike works

Hike is an incredibly simple service for the user to use. After signing up for an account with us, they simply have to apply to companies with their new email - their username@hikemail.net email! Now, the entire job application process is abstracted away from the user - using a combination of the GPT 3.5 model, Levenshtein's distance formula, and Soundex indexing, hike keeps track of every single application you have, as well as the state that the application is on as those further emails coming rolling in.

What it does

The categories include, under review, screening, interview, acceptance or rejection. All of these emails are organized by category as a part of the Hike dashboard, and all emails that are sent to the hikemail.net can also be forwarded to a primary email.

How we built it

The first step in creating hike was developing an email client. This was a very tricky challenge, as it required significant networking work. However, we were able to do this ultimately by using AWS SES, and we store the actual email content in AWS S3. Then, we must dispatch to several lambda functions in order to accomplish both our forwarding and our analysis. One lambda function forwards the emails to the users real email address for their convenience, while the other performs the analysis needed. This analysis is performed by a GPT model that provides the following information about the emails: the company that the email was sent from, the position that was applied to, and the status of the email. Then, this lambda function interfaces with a microservice that we have written in Hono, which does crucial things like Levenshteins distance and Soundex caching in order to figure out if this email is associated with an application already in the database, as in that case we need to figure out if the user has advanced in the application process or not. Finally, the end user will see all of the content they need in a react website, collected by applications.

Challenges we ran into

Handling a project of such large scale was difficult at times and we struggled to make a conducive frontend and backend. There were many moving parts but we stayed patient and worked hard to debug all the problems that we had. Splitting up work also helped us a lot, as well as discussing our issues with each other when we had them.

Accomplishments that we're proud of

We are proud that we constructed such a large project in such little time, as we got the backend functionality and a aesthetically designed frontend that is easy to navigate and interesting to view. We worked with a lot of difficult technologies so we learned a lot in the process, and we are proud of what we ended up building.

What we learned

We learned how to work together as a team to develop a full stack application, as well as rotating between frontend and backend features. We also worked a lot with git branches, which was an interesting experience as we were able to experience how corporations have their software engineers program in order to ensure minimal issues.

What's next for hike

We really do see the potential in this idea. We imagine extensibility in terms of giving users the ability to auto reply to certain types of emails, manage interview scheduling for users. We hope that, if this were to become a good enough product, that hike would become a standard for the job application space.