HiredBy.me

The AI powered chatbot finds the best role for applicants, and best applicants for a role in a friendly way.

Comment

Inspiration

We were frustrated by many companies' application processes. Currently, it's hard to figure out the best role to apply to; resumes aren't able to reflect all of an applicant's potential; often there's no interaction between actual people within the company and applicants: people just get rejected by ATS before any personal interaction happens; and recruiters get too many applications for a role from people who would fit another one better. These pain points inspired us to create a chatbot that can completely revolutionize the way online applications are handled.

What it does

The HiredBy.me bot is able to create a personal interaction with each applicant right on a company's careers page to narrow down what would be the best role for each applicant. The chatbot nonintrusively interacts with applicants, asking them questions that can narrow down a person's preference, as well as determine their level of proficiency in the field that they're interested in, to determine what the best role(s) for them to apply to.

How we built it

We are using two servers, one for the careers application that the applicants and employers use for the application process. The careers application has a list of all roles (jobs), as well as applicants (along with their characteristics determined by the chatbot) and how closely each applicant fits each role. On the second server on Azure, we have all the logic for the chatbot-- how it decides what questions to ask, and conversations to follow. The chatbot logic also has hooks back into the career server to aggregate applicant info into our MongoDB instance. We had to determine what interactions would lead to good insights into each candidate. We used a similar structure to the personality test questionairre. So for instance we would ask if someone knows a programming language, or what their favorite is, and what they would do in tough situations.

Challenges we ran into

So many! We had a very difficult time setting up the chatbot. There were connection issues and environment issues all the way until the morning of finally submitting. It took a lot of time to actually figure out how the bot logic worked. Our CloudDB instance was not playing well with our application. It took a lot of time to really figure out how to set the project up well. We each were totally new to the different parts that we worked on-- but we were still able to divide the project up where we could each complete our parts in time. The primary difficulty we faced in creating our application was coordinating our different skillsets and learning a development stack that was at least somewhat new to each of us. Although the role assignments that we chose proved effective, we were nonetheless slowed by environment problems such as coordinating development between platforms.

Accomplishments that we're proud of

Overall, our team is proud of developing a website that successfully allows the user to have an interactive experience with the "chatbot" that can determine which position the user is best suited for. Additionally, we are proud of how smoothly the back-end and front-end aspects of the project merged to enhance the user experience without a complex visual appearance. We also got a compliment from a Microsoft employee for our shell-fu.

What we learned

We all have very limited experience with each of the areas that we worked in for this project. This was Lauren's first Hackathon, as a college freshman, so practically everything was a new experience for her. She worked on front-end and learned the responsibilities with GitHub, DevPost, and setting up a chatbot. Taslim worked with Microsoft utilzing their API's including the chatbot API. Luke worked on back-end using handelbars and mongodb. Pranathi worked on databases and data collection.

What's next for HiredBy.me

We want to add much better questions and measures of fitness. We also want to add the ability to add new jobs and choose questions from a premade list that can relate to the new positions. We also believe that we can design a neural net that will be trained on all the applications and who's hired as well as how well the hired persons do in job reviews to optimize our chatbot's ability to determine the best fit candidate. This technology is fairly straightforward, many people want something like this, but very few people are actually implementing.

Built With

Share this project:

Updates